Rockwood & Green’s Fractures in Adults
6th Edition

Chapter 38
Injuries of the Craniocervical Junction
Carlo Bellabarba
Sohail K. Mirza
Jens. R. Chapman
Injuries to the craniocervical junction are common, and they are among the few skeletal injuries that carry a high likelihood of death (1,2,3). They are often difficult to diagnose on initial imaging studies. Successful management of these injuries depends on familiarity with the normal anatomic relationships of this region of the spine and recognition of the critical consequences of injured structures.
The craniocervical junction refers to the osseoligamentous and neurovascular structures that extend from the skull base to C2. It is comprised of the highly specialized bony articulations between the occipital condyles, C1 and C2, and the complex ligamentous system linking these three bones into one functional joint. The craniocervical articulation is a very mobile transitional region of the vertebral column. As with other transitional regions of the spine, it is highly susceptible to injury. This region’s vulnerability to injury is particularly high because of the large lever-arm induced rostrally by the cranium and the relative freedom of movement of the craniocervical junction, which relies disproportionately on ligamentous structures rather than on intrinsic bony stability.
Due to the vital functions of the nearby neurovascular structures, injuries to the upper cervical spine that disrupt its structural integrity carry a high likelihood of death. Improved trauma care, and perhaps the availability of airbags and enforcement of seat-belt laws, however, has increased the likelihood of survival in patients with these injuries, raising the burden of responsibility
P.1436

to promptly identify and appropriately treat these life-threatening injuries.
This chapter will focus on the evaluation and treatment of six injury types, which often coexist: (i) occipital condyle fractures, (ii) craniocervical dissociation, (iii) fractures of the atlas, (iv) C1–C2 ligamentous instability, (v) odontoid fractures, and (vi) traumatic spondylolistheses (hangman’s fractures) of C2.
ANATOMY AND BIOMECHANICS
Normal Functional Anatomy
The skull base, atlas, and axis comprise the three bony components of the upper cervical spine and form an integrated functional unit (Fig. 38-1). The five unconstrained joints of the upper cervical spine rely primarily on an intact, multilayered ligamentous system for stability. This unique anatomical arrangement allows the upper cervical spine to contribute a substantial portion of neck motion.
The specific arrangement of ligaments at the craniocervical junction uses the atlas as a washer or base for a coupled, multiplanar motion. The principal structural ligaments bypass the atlas, extending from the bony elements of the basion at the skull base to the odontoid process of C2.
The craniocervical junction accounts for approximately 60% of axial-plane cervical spine rotation, 40% of sagittal-plane flexion-extension motion, and 45% of overall neck motion (4). The normal axial plane C1–C2 rotational excursion ranges from 80 to 88 degrees. The C0–C1 and C1–C2 flexion-extension excursion is 20 to 30 degrees at each level. Total left to right lateral bending amounts to 20 degrees at the C1–C2 segment and 5 to 10 degrees at C0–C1 (5,6,7,8,9). This degree of motion requires a joint architecture that has relatively few bony constraints and therefore strong ligamentous support, which prevents excessive motion and structural failure. Integrity of these ligaments is particularly critical in this area, which contains highly vulner-able neurovascular structures.
FIGURE 38-1 Functional ligament anatomy of the upper cervical spine. The excursions of the three key ligaments of the upper cervical spine are shown in this schematic diagram (shaded areas). (Courtesy of Fred Mann, MD, Professor of Radiology, University of Washington, Seattle.)
The alar ligaments are key to protecting normal craniocervical stability. They are the primary restraints to rotation of the upper cervical motion unit. With an average in vitro load to failure of 210 N, however, these vitally important ligaments tolerate less than 50% load to failure than the cruciate ligaments of the knee (4). These complex ligaments serve a variety of functions. At midposition of the head they are slack. By turning the head in one direction, the alar ligament contralateral to the direction of rotation tightens, while the ipsilateral ligament slackens. Together with the tectorial membrane, the alar ligaments limit flexion. However, they play no role in limiting extension. The contralateral alar ligament limits lateral bending. By sectioning one alar ligament in a cadaveric model, a 30% to 40% increase in flexion, rotation, and lateral bending has been observed (5).
The cranial portions of the anterior and posterior longitudinal ligaments also serve as important ligamentous stabilizers of the craniocervical junction, as do, to a lesser extent, the joint capsules of the respective articulations. The broad tectorial membrane, which constitutes the rostral extension of the posterior longitudinal ligament, effectively limits axial distraction and atlanto-occipital flexion, and is considered, along with the alar ligaments, to be one of the major stabilizing ligaments of the craniocervical junction. Anteriorly, the well-developed atlanto-occipital membrane, an extension of the anterior longitudinal ligament, limits extension, with the thinner anterior atlanto-axial membrane contributing to a less significant degree.
Supplemental ligamentous support of the craniocervical junction is provided by a number of smaller ligaments, such as the apical and cruciate ligaments, the obliquely aligned accessory atlanto-axial ligaments, the anterior atlanto-dental ligament, and the facet joint capsules.
The primary stabilizing ligament of the atlantoaxial motion unit is the cruciate ligament complex with contains the transverse atlantal ligament (TAL). Its in vitro load to failure is350 N (4). By crossing the odontoid at its waist, atlanto-axial flexion, translation, and distraction are minimized, yet rotation is allowed. In flexion, the cruciate ligament is placed under tension, thereby preventing the odontoid from compressingthe spinal cord.
Pathological Conditions
Postmortem studies have shown fracture-dislocations of the craniocervical junction to be a leading cause of death in motor vehicle collisions (2,3,10).
The atlas, which will fracture with as little as 1 to 2 mm of deformation, is the most fragile vertebral segment in humans. It is especially susceptible to bursting-type fractures with relatively low axial loads. Transverse atlantal ligament integrity played no significant factor in the amount of load to failure tolerated by the atlas ring in a cadaveric study (11). Due to its propensity for fracture, the atlas ring serves as a red flag, indicating the potential presence of spine injury elsewhere. The two most vulnerable bony structures of the axis are the pars interarticularis
P.1437

and the odontoid waist. Forced hyperextension can lead to failure of either of these structures. Flexion is believed to be the cause of 80% of odontoid fractures, as the transverse atlantal ligament is forced against the odontoid (12).
Ligamentous instability of the craniocervical junction, even when subtle radiographically, can have grave consequences if left undetected. Intact atlanto-occipitocervical ligamentous structures usually do not allow more than 5 degrees of rotation and more than 2 mm of diastasis (4). The alar ligaments are most susceptible to traumatic rupture with the head in a flexed and rotated position. Atlanto-axial rotation of more than 50 degrees in either direction, as measured by computed tomography (CT) scan is suspicious for alar ligament insufficiency. Greater than 56 degrees of rotation is diagnostic of disruption (13). An intact transverse atlantal ligament limits anterior subluxation of the atlas relative to the axis to 3 mm in adults and 5 mm in children (4). Similarly, atlantoaxial flexion of greater than 5 degrees is highly indicative of transverse ligament insufficiency (4). Atlanto-axial translation of greater than 9 mm in adults indicates a comprehensive failure of all key craniocervical ligaments.
DIAGNOSTIC WORK-UP AND ASSESSMENT
The frailty of the upper cervical spine’s bony structures and its reliance on ligaments for stability renders this region vulnerable to injury. The potential for upper cervical instability should therefore be considered in all patients who have sustained high-energy injuries. In cognitively unimpaired patients, upper cervical spine fractures and dislocations are usually accompanied by neck pain, headache, and nuchal tenderness. Neurologic assessment should be performed according to American Spinal Injury Association guidelines (14). Cranial nerve function should be part of any examination of patients with possible head or neck injuries. The abducens and hypoglossal nerves are most commonly affected by craniocervical injuries. A wide variety of neurologic injury patterns are possible in patients with trauma to the upper cervical spine. These range from complete pentaplegia to incomplete injuries, such as cervicomedullary syndromes and disorders affecting brainstem function. The cervicomedullary syndromes, which include cruciate paralysis as described by Bell and hemiplegia cruciata initially described by Wallenberg, represent the more unusual forms of incomplete spinal cord injury and are a result of the specific anatomy of the spinal tracts at the junction of the brainstem and spinal cord (15). Cruciate paralysis can be similar to a central cord syndrome, although it normally affects proximal more than distal upper extremity function. Hemiplegia cruciata is associated with ipsilateral arm and contralateral leg weakness.
Obtunded patients present increased diagnostic challenges and have to be assessed for possible spinal column and cord injury. In these patients, physical examination is frequently limited to inspection, palpation, and assessment of spontaneous muscle tone, reflexes, and anal sphincter responses. With cognitively impaired patients, spine clearance depends largely on the correct interpretation of x-rays.
As with other joint injuries, upper cervical spine injuries may present in a partially reduced or subluxed position. Since such conditions are much more difficult to identify, clinicians should look for indirect signs of a severe craniocervical disruption, such as hemorrhage and soft tissue swelling, during clinical and x-ray examination.
Plain X-Rays
Plain x-ray evaluation of the cervical spine generally consists of an initial screening cross-table lateral x-ray, followed by the remainder of a routine cervical spine x-ray series, which includes two oblique views, an anteroposterior open-mouth odontoid view, one lower cervical spine anteroposterior x-ray, which usually allows visualization of C4 to T4, and a swimmer’s view.
With regard to the upper cervical spine, the lateral x-ray has several shortcomings (16). Since the typical lateral cervical spine x-ray is centered in the mid-neck region, interpretation of the craniocervical junction can be impaired by parallax or the obliquity of the C1 superior articular surfaces and the occipital condyles. Even minor malrotation of the head further distorts the occipital condyles and the neural arch of the axis, thus reducing the diagnostic value of such x-rays (17).
Certain bony structures of the craniocervical junction should be identified in order to rule out injury. The basion, opisthion, occipital condyles, mastoid processes, tip of the dens, and entire outline of atlas and axis should be visualized and assessed for fracture, rotation, and displacement. The occipito-cervical and atlanto-axial joints should be assessed for congruence. In adults, widening of the prevertebral soft tissue mass in the upper neck is an important warning sign of significant underlying trauma (Fig. 38-2). Established parameters that can help in identifying upper cervical spine trauma are the anterior spinal laminar line, Wackenheim’s line, the atlanto-dens interval (ADI) and the space available for the cord (18). The anterior spinal laminar line can be drawn between the opisthion and the anterior cortex of the posterior arch of the atlas and laminae of the axis and C3. In the absence of fracture or dislocation all these reference points should fall within 1 to 2 mm of this line. Wackenheim’s line is determined by drawing a straight line as a posterior caudal projection of the clivus, directed toward the upper cervical spine. The tip of the odontoid should be within 1 to 2 mm of this line. In adults the anterior cortex of the odontoid should fall within 3 mm of the posterior cortex of the anterior arch of C1 (ADI). The anterior odontoid cortex should also parallel the posterior cortex of the atlas. The space available for the cord should measure at least 13 mm in adults. Transverse atlantal ligament insufficiency should be suspected if there is an atlanto-dens interval above 3 to 5 mm in adults or with the loss of atlanto-dental parallelism. Deviation from these expected reference parameters should prompt further scrutiny (17,19,20,21,22).
P.1438

Anteroposterior visualization of the upper cervical spine requires either an open-mouth odontoid view or coronal CT reconstruction (16). Normal studies are expected to demonstrate symmetrically articulating occipital condyles, C1 lateral masses, and C2 superior articular processes. The odontoid should be well centered between the lateral masses of C1. There should be no overhang, translation, or distraction between the lateral masses of C1 and C2. The articular surfaces of occipital condyles, the C1 superior facets and the components of the atlanto-axial articular surfaces should be equidistant to one another.
FIGURE 38-2 A. Prevertebral soft tissue shadow. In a healthy recumbent adult without endotracheal tube the prevertebral soft tissue shadow should not exceed 6 mm. B. Bony screening lines and dens angulation. The anterior cortex of the odontoid should parallel the posterior cortex of the anterior ring of the atlas. Any kyphotic or lordotic deviation should be viewed with suspicion for an odontoid fracture or transverse atlantal ligament (TAL) disruption. Wackenheim’s line is drawn as a continuation from the clivus caudally. The tip of the odontoid should be within 1 to 2 mm of this line. The C1–C3 spinolaminar line’s reference points are drawn from the anterior cortex of the laminae of the atlas, axis, and C3 segments, which should fall within 2 mm of one another. Greater deviation should raise suspicion of atlantoaxial translation or disruption of the neural arches of either segment. C. Ligamentous injury reference lines (lateral x-rays). The atlas-dens interval (ADI) should be less than 3 mm in an adult (5 mm in a child). The space available for the cord (SAC) is measured as the distance from the posterior cortex of the odontoid tip to the anterior cortex of the posterior arch of the atlas and should amount to more than 13 mm. The dens-basion interval (DBI) is the distance between the odontoid tip and the distal end of the basion. It should be less than 12 mm in adults. The posterior axis line (PAL-B) should not be more than 4 mm anterior and should be less than 12 mm posterior to the basion. D. Bony screening lines (anteroposterior imaging). The left and right lateral atlas-dens intervals (LADIs) should be symmetric to one another (with 2 mm deviation). The bony components of the atlanto-occipital joints should be symmetric and should not be spaced more than 2 mm apart on anteroposterior images. (Courtesy of Fred Mann, MD, Professor of Radiology, University of Washington, Seattle.)
A variety of x-ray reference lines have been suggested as screening tools for occipitocervical ligamentous injuries. Powers suggested calculating the ratio of the distance from the basion (A) to the anterior spinal laminar line (C), to the distance from the posterior cortex of the atlas (B) to the opisthion (D). Powers’ ratio, which is limited to screening for anterior occipito-cervical dislocations, is considered abnormal if the ratio of AC to BD is less than 0.8 (23). The “X-lines,” as described by Lee, are a series of intersecting reference lines and their relationship to common landmarks such as the odontoid tip (24). Unfortunately,
P.1439

Powers’ ratio and the X-line system are both hampered by complexity and the lack of interobserver reliability. A simple and helpful screening method for identifying occipitocervical injury was described by Harris et al (Fig. 38-2C) (17,25). In adults, the distance between the basion to the upper tip of the odontoid (basion dens interval, or BDI) is less than 12 mm in 95% of individuals. The basion lies anywhere between 4 mm anterior to 12 mm posterior to a line projecting cranially from the posterior cortex of the C2 vertebral body (basion axis interval, or BAI) in 98% of the population. In a prospective comparison study, a significantly higher sensitivity and specificity was found using the BAI and BDI compared to the Power’s ratio and Lee’s X-lines (17,25). Any deviation from the normal range of values described above warrants further evaluation of the craniocervical junction.
Anteroposterior visualization of the occipitocervical junction is possible with open mouth odontoid views or coronally reformatted CT of the upper cervical spine (Fig. 38-2B) (16,18). Reference points on the anteroposterior x-rays are the occipital condyles, the lateral masses of the atlas, and the odontoid with the lateral masses of the axis. Any odontoid asymmetry relative to the lateral masses of the atlas or any diastasis of the upper cervical spine articulations suggests a ligamentous injury. Provocative standard x-rays for instability, such as flexion-extension x-rays, are discouraged in patients with suspected occipitocervical dissociation because of the associated potential for neurologic compromise to the spinal cord (26). Cranial traction applied under lateral fluoroscopic monitoring by a surgeon with expertise in this area of the spine can be helpful in assessing for borderline cases with unclear craniocervical stability (27,28,29,30,31).
The timing of dynamic motion x-rays of the neck continues to be a source of controversy. Flexion-extension x-rays are helpful in eliciting less than obvious instability of the neck and assessing the healing results of the cervical spine. However, they are usually contraindicated for patients with known acute cervical spine fractures because of the risk of neurologic worsening. It remains our preference to defer any motion studies until a patient has documented absence of cognitive impairment or has decreased neck tenderness (32,33). Cognitively unimpaired patients who can follow physician instructions, are not under the influence of mind-altering substances, are neurologically normal and have no x-ray evidence of fracture or dislocation can be considered for flexion-extension x-rays to evaluate for more subtle degrees of instability. In obtunded patients, physician conducted fluoroscopic flexion-extension views have been proposed as a means of clearing the cervical spine (27,28,29).
Because of the unique anatomy of the upper cervical spine and its structural reliance on the alar ligaments, fluoroscopically guided traction testing should be considered prior to performing a flexion-extension maneuver. The patient is positioned supine on a fluoroscopy table with folded sheets under the shoulder girdle. With the C-arm positioned in a lateral trajectory centered on the arch of C-1, a “neutral” baseline image is obtained and stored for comparison. Using a 5-lb weight applied by head-halter or cranial tong traction, the image is repeated and immediately assessed for any distractive changes between the occiput, atlas, and axis. Should there be no appreciable distraction between these structures, the traction is increased to 10 lb and compared to the baseline image. If there is no appreciable distractive upper cervical spine instability, the physician may choose to assess the upper cervical spine in flexion and extension under fluoroscopic guidance (31). As with the distraction test, this examination is discontinued should a fracture or distraction of more than 2 mm between any of the bony elements be identified. In healthy adults, the atlanto-occipital joints should not distract more than 2 mm, and the atlanto-axial joints no more than 3 mm (34). After assessing upper cervical spine stability the examining physician can proceed with fluoroscopic evaluation of the lower cervical spine.
Computed Tomography
Computed tomography (CT) is indicated in the assessment of patients with known or suspected cervical spine fractures and is a crucial diagnostic tool for any area of the spine that is not sufficiently visualized on plain x-rays. With conventional 3 or 5 mm axial cut thickness, it can be relatively easy to miss axial plane fractures such as odontoid fractures or occipito-cervical and atlanto-axial dissociations (35). Similarly, sagittal and coronal reformations based on fine-cut scans with slices no more than 2 mm wide are helpful in better understanding articular incongruities or complex fracture patterns and in improving visualization of transversely oriented fractures (16,35).
A head CT scan that incorporates the craniovertebral junction can be helpful in detecting occipital condyle fractures that are commonly missed on conventional x-rays (36). They are also useful in detecting subarachnoid craniocervical junction hematomas that may be associated with atlanto-occipital dissociation (37). Therefore, including the inferior margin of the foramen magnum in all head CT scans with bone and soft tissue images appear to be an important imaging strategy for patients with substantial head trauma (38).
In the past, CT scans have largely played a supplemental role to plain x-rays in the work-up of cervical spine trauma. With the advent of rapid acquisition helical scanning techniques this traditional approach is being reevaluated. CT scans of the head have become a priority for all patients with suspected or evident head injuries after initial stabilization. The addition of a full cervical spine CT screen using a helical scanning protocol adds minimal imaging time to a head CT, yet provides the clinician with a full set of cervical spine images from the occiput to T-4 including reformatted views of the upper and lower cervical spine.
Three-dimensional image reformations, obtained from fine-cut CT scans, are rarely clinically useful but may assist with the interpretation of more unusual upper cervical injury patterns (Fig. 38-3). Postmyelography CT scanning may be indicated in specific clinical circumstances, mainly if magnetic resonance imaging (MRI) is either unavailable or contraindicated.
FIGURE 38-3 C1–C2 atlantoaxial fracture dislocation. This 48-year-old man sustained isolated neck injuries after a retaining wall collapsed. There was no identifiable neurologic injury. A. A lateral cervical spine x-ray revealed an abnormal atlas and odontoid profile and a minimally displaced C7–T1 facet fracture dislocation (not depicted here). B. Axial CT scan shows a rotational displacement of the atlas relative to the axis. C. Coronal reformatted views highlighted a type II odontoid fracture and lateral atlantoaxial subluxation. For patients with rotatory upper cervical spine problems, three-dimensional CT scan reformatted views can occasionally be helpful. D. A posterior view gives an impression of the rotatory deformity of this fracture. It also demonstrates the neural canal to be relatively well preserved. E. A lateral view reveals the uncovered superior facet of the axis. F. This fracture dislocation reduced with 20 lb traction applied to the patient by Gardner-Wells tongs. Postoperative lateral (G) and anteroposterior (H) x-rays show transarticular atlantoaxial screw fixation with the modified Gallie technique as treatment for the upper cervical spine injury and pedicle screw fixation with rod instrumentation for the lower cervical spine fracture. The patient has had an uneventful postoperative course.
P.1440

P.1441

Magnetic Resonance Imaging
Magnetic resonance imaging (MRI) of the craniovertebral junction is indicated for patients with spinal cord injury and can be helpful in assessing upper cervical spine ligamentous injuries as well as subarachnoid and prevertebral hemorrhage (Fig. 38-4). T2-weighted fat suppression images can demonstrate disruption of ligaments normally responsible for upper cervical stability (38,39). Transverse atlantal ligament injuries can be diagnosed with MRI but are not always obvious (40,41).
Nuclear Imaging Tests
Radioisotope-based imaging tests, such as the technetium-99 bone scan, are rarely necessary in the work-up of acute upper cervical spine trauma. This test can, however, be useful in assessing patients for subclinical fractures of the upper cervical spine. A technetium-99 bone scan would be expected to show an abnormal uptake in the injury zone as early as 3 days after the injury (42). For patients with a history of chronic neck pain or headaches, a technetium-99 scan with single photon emission computed tomography enhancement can be helpful in identifying osteoarthritis of the articulations of the upper cervical spine.
Electrodiagnostic Tests
There are few indications for obtaining electrodiagnostic tests in patients with upper cervical spine fractures. Dermatomal somatosensory evoked potentials can help in understanding the neurologic injury pattern in patients with upper cervical radiculopathy or unusual spinal cord injury presentation. The cranial extent of electromyography utilization is limited to a C4 level if phrenic stimulation is used. As in any patient who is not examinable due to mental status changes and is suspected of having sustained a spinal cord injury, baseline somatosensory and motor evoked potentials may indicate the presence of neurologic deficits.
Specific Injuries and Their Diagnostic Work-Up
Occipital Condyle Fractures
With the advent of routine head CT scans in the assessment of head injuries, fractures to the occipital condyles are being more consistently recognized (Fig. 38-5). An open mouth odontoid view may also demonstrate a fracture of the condyles. Fractures of this region are best visualized using a slice thickness of 2 mm or less, with sagittal and coronal reformations (20,43). An accurate assessment of stability and the recognition of craniocervical dissociation are the key diagnostic determinations. Should the stability of the occipital condyle injury be in question a provocative fluoroscopic evaluation may be helpful (27,28,29,30,31).
Craniocervical Dissociation
Due to their widespread availability, plain x-rays remain the most important screening test for this rare but survivable injury.
P.1442

P.1443

P.1444

Prevertebral soft tissue swelling and distracted or incongruous occipitocervical articulations are important findings. Useful screening tools are the Harris lines, Powers’ ratio, and Wackenheim’s line (Fig. 38-2). Patients with suspected or established atlanto-occipital dissociation should receive a fine-cut CT scan, with reformations, of the craniovertebral junction as well as an MRI scan, if available (44). Patients with unclear stability at the craniocervical junction are candidates for provocative traction testing of the cervical spine under fluoroscopic visualization (Fig. 38-6) (27,28,29,31).
FIGURE 38-4 Since the major craniocervical stabilizing ligaments extend from the occiput to C2, the atlanto-occipital and atlanto-axial joints should be considered as a single functional unit, as demonstrated in this example of craniocervical dissociation affecting both the C0–C1 and C1–C2 articulations in a woman involved in a motor vehicle collision. A. Parasagittal CT scan through the C0–C1 and C1–C2 joints shows incongruence and distraction across both joints. B. Accordingly, parasagittal T2 MRI images show increased signal intensity within both these same two articulations. C. Midline sagittal CT shows a type I odontoid fracture (black arrow), a common finding in craniocervical distractive injuries that represents alar ligament avulsion. White arrows demonstrate an atypical atlas fracture representing tension failure of the anterior C1 ring. D. Midline sagittal MRI image shows several other typical features of craniocervical dissociation, including extensive retropharyngeal hemorrhage (white arrows), tectorial membrane rupture (black arrow), and rupture of the posterior ligaments between the C1–C2 (gray arrows). E. The patient had a progressive spinal cord injury and was taken emergently for occiput to C3 stabilization, after which her neurologic status gradually improved.
FIGURE 38-5 A,B. Lateral and open-mouth odontoid views of a 43-year-old laborer who fell from a 40-foot scaffold. The patient complained of severe neck pain on arrival at the hospital, but was found to be neurologically normal. The lateral x-ray reveals increased prevertebral swelling. C–E. Axial computed tomography (CT) of the foramen magnum and reformatted CT reveals type III occipital condyle fracture. The axial CT of the atlas revealed an anterior arch fracture. F. The T2-weighted sagittal magnetic resonance imaging (MRI) shows high intensity signal changes around the upper C-spine. G,H. These lateral x-rays demonstrate pathological distraction at the craniocervical junction with 5 lb of cranial tong traction. This otherwise occult occipitocervical dislocation was treated with initial closed reduction using a halo-vest, followed by open reduction and instrumented occipitocervical arthrodesis 3 days later.
Fractures of the Atlas
Fractures of the anterior and posterior arch are usually visible on lateral x-rays. The assessment of atlanto-axial lateral mass
P.1445

congruence on an open mouth x-ray (45) or coronal CT reformation carries considerable therapeutic value. Lateral displacement of the C1 lateral mass lateral to the outer cortex of the C2 lateral mass raises concerns for the structural integrity of the transverse atlantal ligament (see Indications For Surgical Treatment) (Fig. 38-7). CT scanning provides the best method of evaluating important atlas fracture characteristics.
FIGURE 38-6 Provocative traction x-rays for staging of craniocervical instability. A. Lateral cervical spine fluoroscopic view of a patient involved in a high-speed MVC who presented with neck pain. Minimal (1 mm) subluxation at the C0–C1 articulation and normal basion-dens and basion-axis intervals were noted on CT scan (not shown). MRI scan showed increased signal intensity at the C0–C1 and C1–C2 joints (not shown). B. Lateral fluoroscopic view with 5 lb of cranial tong traction demonstrates greater than 2 mm of widening across the C0–C1 joints. This positive provocative traction test confirms a highly unstable distractive craniocervical ligamentous injury that requires operative stabilization. A craniocervical dissociation that presents with deceptively minimal displacement but in which craniocervical instability is confirmed with provocative traction x-rays is defined as type II according to the Harborview classification system of craniocervical injuries (see Table 38-2).
Atlanto-Axial Dislocations
Transverse atlantal ligament injuries in the absence of atlas fractures can be difficult to diagnose on plain x-rays. If the atlanto-dens interval exceeds 3 mm in adults or 5 mm in children, or if the ventral cortical surface of the dens is not parallel to that of the posterior cortex of the anterior arch of the atlas, TAL insufficiency should be suspected (Fig. 38-8). If atlanto-axial instability is not apparent on plain x-rays, flexion-extension x-rays can be helpful in determining the presence of dynamic instability. Although visualization of the transverse atlantal ligament on MRI scan has been described, diagnostic consistency remains doubtful for routine use (40,41).
FIGURE 38-7 A. The open-mouth odontoid view shows bilateral overhang of the C1 lateral masses relative to the C2 facets, with combined lateral displacement measuring 13 mm. B. Axial CT image shows a true Jefferson fracture in the form of a four-part burst fracture of the atlas. This fracture is unstable due to associated transverse atlantal ligament disruption.
Asymmetry of the C1 lateral masses with respect to the odontoid and the C2 lateral masses on open mouth anteroposterior x-rays suggests the presence of atlanto-axial rotatory subluxation (46). Quantification of the deformity can be best performed with fine-cut CT scanning (47,48) and may require dynamic views.
P.1446

Atlanto-axial dissociation can be identified on lateral screening cervical spine x-rays by the lack of overlap between the odontoid tip and the C-1 ring (Fig. 38-9). Open-mouth odontoid x-rays and CT scans with coronally reformatted views provide additional insight into the rotatory deformity (49).
Odontoid Fractures
Plain x-rays show the majority of type I and II odontoid fractures and virtually all type III injury subtypes. Minimally displaced fractures can go unrecognized on lateral cervical spine x-rays due to cortical overlap of the convex superior articular surfaces of the axis (50). Axial CT images may insufficiently visualize type II injuries if the study’s axial image thickness is too wide or if the axial image extends through the fracture zone (35). Sagittally and coronally reformatted views are therefore important diagnostic tools in the assessment of a potential odontoid fracture.
FIGURE 38-8 Translational (type B) atlanto-axial subluxation with bony (type II) transverse atlantal ligament avulsion in a patient involved in a high-speed MVC. A. The lateral cervical spine x-ray shows widening of the atlanto-dens interval and angulation between the posterior cortex of the C1 arch and anterior cortex of the odontoid process (black lines). Both of these findings are suggestive of acute transverse atlantal ligament incompetence. B. Axial CT image shows avulsion fracture at the left TAL insertion (white arrow). C. Lateral cervical spine x-ray 6 months after open reduction and posterior instrumented C1–C2 arthrodesis. Because of aberrant vertebral artery anatomy, fixation was achieved with a C1 lateral mass and C2 pars interarticularis screw construct rather than transarticular screws.
Traumatic Spondylolistheses of the Axis
Hangman’s type fractures can usually be identified on lateral plain x-rays. Fracture of the neural arch can involve the posterior
P.1447

aspect of the C2 vertebral body, the C2–C3 facet joint, and or any part of the intervening neural arch. Disruption of the posterior spinal laminar line due to posterior expansion of the neural canal can be seen in more displaced neural arch fractures. Disco-ligamentous stability may have to be assessed in some of the more displaced fractures. This is best accomplished with a traction test under lateral fluoroscopic visualization (51).
FIGURE 38-9 Distractive (type C) atlantoaxial instability. A. Lateral cervical spine x-ray showing severe cranio-cervical distractive injury, with displacement primarily through the C1–C2 joint. Note the absence of overlap between the atlas and the odontoid process (black lines). B. Sagittal CT image through the C0–C1 and C1–C2 articulations shows wide displacement across the C1–C2 articulation. Note the subtle associated C0–C1 anterior subluxation. Since the major craniocervical ligamentous stabilizers—the alar ligaments and tectorial ligament—extend from the foramen magnum to C2, distractive injuries at either of these two joints frequently result in instability at the adjacent craniocervical articulation, which must be carefully evaluated to determine the necessary extent of fixation.
INJURY CLASSIFICATION
Occipital Condyle Fractures
Occipital condyle fractures may be highly unstable if they represent bony avulsion of major craniocervical stabilizers. Autopsy studies by Saternus suggested six possible occipital condyle fracture variants, resulting from force vectors along six different axes (52). Anderson subsequently described a condensed version of this classification system (Table 38-1) (Fig. 38-10) consisting of three categories. Type I fractures are usually stable, comminuted axial loading injuries. Type II injuries are potentially unstable
P.1448

injuries caused by a shear mechanism that results in an oblique fracture extending from the condyle into the skull base. Type III injuries are unstable avulsion injuries that result in a transverse fracture line through the occipital condyle (Fig. 38-11) (53). Any occipital condyle fracture should be considered a possible component of craniocervical dissociation (Fig. 38-1).
TABLE 38-1 Classification of Occipital Condyle Fractures
Injury Type Distinguishing Characteristics Significance
I Comminuted fracture of an occipital condyle Stable injury treated with a cervical collar, possibly a halo for severe collapse, unless associated with craniocervical dissociation
II Extension of a basilar skull fracture into an occipital condyle Stable injury treated with a cervical collar unless associated with craniocervical dissociation
III Avulsion fracture at alar ligament insertion Labeled unstable in the original description, but commonly treated with halo unless associated with craniocervical dissociation
FIGURE 38-10 Anderson and Montesano classification of occipital condyle fractures. A. Type I injuries are comminuted, usually stable, impaction fractures caused by axial loading. B. Type II injuries are impaction or shear fractures extending into the base of the skull, and are usually stable.C. Type III injuries are alar ligament avulsion fractures and are likely to be unstable distraction injuries of the craniocervical junction.
Craniocervical Dissociation
Poor x-ray visualization and incomplete understanding of the ligamentous anatomy may present challenges in assessing cranio-cervical stability. Obvious signs of instability are translation or distraction of more than 2 mm in any plane (4), neurologic injury, or concomitant cerebrovascular trauma (54). Traynelis has identified three craniocervical dissociation patterns according to the direction of displacement of the cranium relative to the cervical spine (55). Such directional classification systems have two major limitations. First, the extreme instability of these injuries renders the position of the head relative to the neck completely arbitrary and more dependent on external forces at the time of imaging than on any intrinsic injury characteristic (30). The direction of x-ray displacement therefore has little influence on prognosis or treatment method. Second, they do not reflect injury severity or the potential for spontaneously reduced dislocations, causing the magnitude of displacement to potentially underestimate the degree of instability.
A different classification system has been proposed that attempts to quantify the stability of the craniocervical junction. The problem lies in segregating patients with minimally displaced (<2 mm) craniocervical injuries into those with relatively stable injuries who can be treated nonoperatively and those with highly unstable but partially reduced injuries who require operative stabilization in spite of a misleadingly low degree of displacement. The Harborview classification system uses manual traction to categorize patients with minimally displaced injuries (<2 mm) into stable (type I) and unstable (type II) groups. Patients with gross instability (>2 mm displacement) are categorized as type III injuries. Surgical stabilization is reserved for patients with types II and III injuries of the craniocervical junction, which are defined as dissociations (Table 38-2).
Fractures of the Atlas
It is most useful to view atlas fractures as either stable or unstable injuries (Table 38-3) (41). Instability invariably equates to the presence of TAL insufficiency, which can be diagnosed either
P.1449

by direct means, such as by identifying bony avulsion on CT scan or ligament rupture on MRI, or indirectly by identifying widening of the lateral masses (Fig. 38-7) with ≥l7 mm lateral overhang relative to the lateral masses of C2 (45), appropriately corrected for x-ray magnification, if applicable (see below) (56,57).
FIGURE 38-11 Type III occipital condyle fracture with craniocervical dissociation. A. Lateral cervical spine x-ray shows dislocation of the atlanto-occipital joints in a patient involved in a high-speed motor vehicle collision who presented with a high cervical spinal cord injury. The white and black arrows indicate the anterior margins of the occipital condyles and C1 lateral masses, respectively. B. Coronal CT image shows an associated avulsion fracture of the left occipital condyle (gray arrow), resulting in functional incompetence of the attached alar ligament. The remainder of the occipital condyle is dislocated and does not articulate with the C1 lateral mass.
TABLE 38-2 Harborview Classification of Craniocervical Injuries
Stage Description of Injury
1 MRI shows hemorrhage or edema at the craniocervical junction.
Craniocervical alignment is normal by Harris lines.
No distraction on traction test with 25 lb of traction.
2* MRI shows hemorrhage or edema at the craniocervical junction.
Craniocervical alignment is normal by Harris lines.
Traction at weights less than 25 lb shows sufficient distraction to meet craniocervical dissociation threshold established by the Harris measurements.
3* Static imaging studies show distraction beyond thresholds of Harris measurements.
* Represent injuries defined as craniocervical dissociation.
Levine and Edwards (58) have described a useful four-part classification system which categorizes atlas fractures into: (i) posterior arch fractures, caused by hyperextension forces, which are stable injuries; (ii) lateral mass fractures, caused by rotation or lateral flexion forces, which are unstable injuries; (iii) isolated anterior arch fractures, which are caused by hyperextension, and have been divided into minimally displaced, comminuted, and unstable injuries (59,60); and (iv) bursting-type fractures (Fig. 38-12) (15,61). As mentioned above, the extent of lateral mass separation is more relevant than the number of fracture fragments.
TABLE 38-3 Classification of C1 Fractures
Injury Type Distinguishing Characteristics Treatment
Stable Posterior arch fracture
Anterior arch avulsion fracture
C1 ring fracture with <7 mm of overall lateral mass displacement
C1 ring fracture with ≥7 mm of overall C1 lateral mass displacement
Rigid collar
Rigid collar
Rigid collar or halo-vest
Traction followed by Halo-vest vs C1–C2 arthrodesis
Unstable Anterior arch fracture with posterior displacement relative to the dens (plough fracture) Halo-vest vs C1–C2 arthrodesis
Atlanto-Axial Injuries
Three atlanto-axial instability patterns may present either as isolated or combined injuries. Type A injuries are rotationally
P.1450

displaced in the transverse plane; type B injuries are translationally unstable in the sagittal plane due to transverse atlantal ligament insufficiency; and type C injuries are characterized by vertical atlanto-axial distraction and represent a variant of cranio-cervical dissociation (Table 38-4).
FIGURE 38-12 Classification of atlas fractures (according to Levine). A. Isolated bony apophysis fracture. B. Isolated posterior arch fracture. C. Isolated anterior arch fracture. D. Comminuted lateral mass fracture and (E) burst fracture, three or more fragments.
Type A injuries (Fig. 38-13): Rotational displacement of the atlanto-axial motion segment is most commonly nontraumatic and will therefore not be described in detail. However, traumatic causes have been described and range in severity from mild rotational subluxation to complete dislocation of the atlanto-axial lateral masses (62). The classification outlined by Fielding remains useful in describing and treating these unusual injuries.
Type B injuries: Translational atlanto-axial instability is the result of transverse atlantal ligament insufficiency. The presence of a purely ligamentous tear (type I) versus a bony avulsion injury (type II) (Fig. 38-8) may be pivotal in determining the need for operative intervention (41,63).
Type C injuries (Fig. 38-9): Distractive atlanto-axial injuries, or atlanto-axial dissociation, constitute a variant of craniocervical dissociation since the disrupted primary ligamentous stabilizers—the alar ligaments and tectorial membrane—extend from C2 to the occiput (19,30,64). They frequently coexist with overt atlanto-occipital distraction injuries (Fig. 38-4) (30,65,66).
TABLE 38-4 Classification of C1–C2 Dislocations
Injury Type Distinguishing Characteristics Significance
A Rotation centered on the dens, where the transverse atlantal ligament is normally intact Treated with closed reduction and immobilization. Beware of associated fractures
B Translation between C1–C2, where transverse ligament is disrupted Mid-substance transverse ligament tears (type I) are treated with C1–C2 arthrodesis
C Distraction indicating craniocervical dissociation Bony avulsions (type II) may be treated with halo or C1–C2 arthrodesis
Treated with arthrodesis and internal fixation
Odontoid Fractures
Fractures of the odontoid process are the most common of axis fractures (41%) (67). All odontoid fractures are considered unstable. Anderson and D’Alonzo’s three-part odontoid fracture classification system has become the basis for odontoid fracture management (Table 38-5) (Fig. 38-14) (68). Type I injuries are considered bony avulsions of the alar ligament from its superolateral odontoid insertion, and may represent a component of craniocervical dissociation. Type II injuries are located at the odontoid waist, in the area covered by the TAL, and have the highest propensity for pseudarthrosis, probably due to their small cross-sectional fracture surfaces and the interruption in blood supply to the cephalad fragment. A type IIA subtype of odontoid fracture has been described by Hadley et al and consists
P.1451

of a highly unstable, segmentally comminuted injury extending from the waist of the odontoid into the body of the axis (69). Type III fractures extend into the cancellous vertebral body and have wider, well-vascularized cancellous fracture surfaces.
FIGURE 38-13 Classification of rotary atlantoaxial subluxation. A. Type I, rotation without translation. B. Type II, unilateral lateral mass subluxation of 3 to 5 mm. C. Type III, unilateral subluxation of greater than 5 mm. Transverse atlantal ligament is ruptured. D. Type IV, posterior displacement of C1–C2.
Variations of the Anderson and D’Alonzo classification system have been proposed as a means of further refining its usefulness. These include differentiation of type II fractures into “high” and “low” odontoid fractures (70), and recent modifications aimed at (i) establishing a precise distinction between types II and III fractures based on the presence or absence of C1–C2 facet involvement and (ii) subdividing type II fractures based on factors specifically deemed to influence management, such as fracture line obliquity, displacement, and comminution (71).
TABLE 38-5 Classification of Dens Fractures
Injury Type Distinguishing Characteristics Significance
I Avulsion injury at insertion of the alar ligament (cephalad to the transverse ligament) Treated with halo, or surgery if associated with craniocervical dissociation.
II Fracture at the waist of the dens where it consists primarily of a ring of cortical bone surrounded by synovial capsule High risk of nonunion. Usually treated with a halo initially, but strong indication for surgery if widely displaced or if fracture shows continued movement despite halo immobilization.
III Fracture extending into cancellous bone within the C2 vertebral body. Treated with a halo or brace.
Traumatic Spondylolisthesis of the Axis
Hangman’s fractures are the second most common type of axis fracture (38%) and are a common type of fatal cervical spine injury (67,72). Effendi has described the following simple classification, as subsequently modified by Levine and Starr (Table 38-6) (Fig. 38-15) (51,73,74).
FIGURE 38-14 The odontoid fracture classification of Anderson and D’Alonzo. A. Type I fractures of the odontoid tip represent represent alar ligament avulsions. B. Type II fractures occur at the odontoid waist, above the C2 lateral masses. C. Type III fractures extend below the odontoid waist to involve the body and lateral masses of C2. Hadley has added the type IIA fracture with segmental comminution at the base of the odontoid (not shown).
P.1452

P.1453

Type I injuries consist of a minimally displaced, relatively stable fracture of the pars interarticularis that results from hyperextension and axial loading. Type IA fractures are atypical unstable lateral bending fractures that are obliquely displaced and usually involve only one pars interarticularis, extending anterior to the pars and into the body on the contralateral side (74). The oblique plane of these fractures makes them less obvious on lateral x-rays, giving the appearance of an elongated pars (Fig. 38-16). Type II fractures are displaced injuries that result when a flexion force follows the initial hyperextension and axial loading insult. Type II injuries may appear similar to type I injuries on supine x-rays but will displace on upright x-rays. Physician-supervised flexion-extension x-rays of type I injuries have also been advocated to differentiate them from spontaneously reduced type II injuries (75). Type IIA injuries are thought to occur from a flexion-distraction mechanism and are more unstable due to their associated C2–C3 disc and posterior longitudinal ligament disruption. Because they are flexion-distraction injuries, kyphosis is the prevailing deformity rather than translation (Fig. 38-17). An inconsistent feature of type IIA injuries is that, because of the injury mechanism, the pars interarticularis fractures tend to be more horizontally oriented than in standard type II injuries. Levine has postulated that any injury where distraction of the C2–C3 disc space occurs with only 10 lb of traction should be considered type IIA. Type III injuries are unusual and highly unstable injuries in which the pars interarticularis fractures are usually minimally displaced and are associated with a complete unilateral or bilateral C2–C3 facet dislocation, which is not generally reducible by nonoperative means. On rare occasions, these injuries may also spontaneously reduce and have the more benign appearance of a type I injury on initial supine lateral x-rays (Fig. 38-18). Of all traumatic spondylolisthesis injuries, type III hangman’s fractures are most commonly associated with spinal cord injury, although displaced atypical (type IA) injuries may result in displacement patterns that result in spinal cord compression and spinal cord injury (Fig. 38-19).
FIGURE 38-15 Classification of traumatic spondylolisthesis of the axis (hangman’s fracture) (according to Effendi, modified by Levine). A. Type I, nondisplaced fracture of the pars interarticularis. B. Type II, displaced fracture of the pars interarticularis. C. Type IIa, displaced fracture of the pars interarticularis with disruption of the C2–C3 discoligamentous complex. D. Type III, dislocation of C2–C3 facets joints with fractured pars interarticularis.
TABLE 38-6 Classification of C2 Arch (Hangman’s) Fractures
Injury Type Distinguishing Characteristics Significance
I Nondisplaced (displacement <2 mm) fracture through the arch of C2 Treated with a collar, occasionally with halo
IA Atypical fracture involving C2 arch on one side and vertebral body on contralateral side. Often extends into vertebral artery foramen Displacement of these atypical fractures may result in considerable canal compromise and spinal cord injury.
Usually treated with a halo. May require reduction and fixation if severely displaced or in the presence of spinal cord injury. Fixation options include C2–C3 anterior arthrodesis, posterior C1–C3 vs C2–C3 arthrodesis, or combined approach (usually reserved for SCI)
II Displaced fracture of C2 arch Treated with a halo
IIA Fracture of C2 arch associated with disruption of the C2–C3 intervertebral disc, showing angulation of the C2–C3 endplates of anterior translation of C2 body on C3 body Treated with halo, and if markedly displaced, possibly direct fixation of fractured arch through a posterior approach, or by C2–C3 anterior arthrodesis
III Fracture of the C2 arch associated with dislocation of the C2–C3 facet joints Frequently associated with neurologic deficit. Requires open reduction and posterior arthrodesis
Other classification systems have evaluated fracture stability
P.1454

P.1455

P.1456

P.1457

based on the degree of translational and angulatory displacement as a measure of the integrity of C2–C3 disco-ligamentous elements (76).
FIGURE 38-16 Type IA traumatic spondylolisthesis of C2. In this atypical hangman’s fracture variant, the fracture lines are “staggered” on the lateral cervical spine x-ray (A), giving the impression of an elongated pars (white arrow). B. Axial CT image shows typical position of pars interarticularis fracture (gray arrow) on the left side, and atypical contralateral fracture extending into the vertebral body and foramen transversarium (black arrow). Greater displacement of the vertebral body fracture at the spinal canal may result in spinal cord compression from the lateral edge of the atypical fracture (black arrow), leading to a higher likelihood of spinal cord injury with type IA injuries than with other type I or II injuries. In this minimally displaced fracture, the patient remained neurologically intact, and was treated successfully with a halo-vest.
FIGURE 38-17 Type IIA traumatic spondylolisthesis of the axis secondary to high speed motor vehicle collision in a septuagenarian. A. Lateral cervical spine x-ray demonstrates the predominance of angulation over translation (black lines), which is pathognomonic for type IIA C2 arch fractures. The causative flexion-distraction mechanism is thought to result in progressive tension failure of the posterior atlantoaxial membrane (white arrow), posterior longitudinal ligament, posterior annulus, and intervertebral disc. The anterior annulus and anterior longitudinal ligament are thought to remain intact. The patient was treated with open reduction and instrumented arthrodesis through a posterior approach. Although interfragmentary screws were placed across the C2 pars interarticularis fractures bilaterally, the instrumentation was extended to C1 in this particular patient due to his advanced age and osteoporosis, as shown on postoperative lateral cervical spine x-rays (B) and midline sagittal CT image
FIGURE 38-18 Delayed diagnosis of type III traumatic spondylolisthesis of the axis. A. Lateral supine cervical spine x-ray showing what appears to be a type I traumatic spondylolisthesis of the axis sustained by a patient involved in a motor vehicle collision. The patient was discharged in a rigid cervical collar. No upright x-rays were obtained prior to discharge. The patient presented for reevaluation with worsening neck pain and the onset of upper extremity paresthesias. B. Upright lateral cervical x-ray in a rigid collar showed the minimally displaced pars interarticularis fracture, but with associated C2–C3 facet dislocation, consistent with a type III traumatic spondylolisthesis of C2. Facet reduction in this highly unstable injury was achieved primarily with supine positioning and assisted by gentle manipulative reduction. Type III hangman’s fractures typically cannot be reduced by closed means due to the dissociation between the body of the axis and the dislocated posterior elements. Because a closed reduction was possible in this patient, a C2–C3 anterior cervical discectomy and instrumented interbody arthrodesis was performed (C), rather than the more typical posterior intervention. Upright post-immobilization x-rays of any type I hangman’s fracture are required to exclude a more unstable injury.
FIGURE 38-19 Type IA traumatic spondylolisthesis of the axis with spinal cord injury. A. Lateral cervical spine x-ray shows displaced traumatic spondylolisthesis of C2 as a result of a motor vehicle collision in a patient who presented with a Brown-Séquard incomplete upper cervical spinal cord injury. B. Axial CT image shows an atypical hangman’s fracture, extending into the body of the axis on the right side. Fracture displacement with this injury pattern causes the posterolateral atypical fracture fragment (white arrow) to encroach on the spinal canal and potentially injure the spinal cord. C. Axial MRI image further demonstrates canal compromise by the posterolateral fracture spike as the likely cause of spinal cord injury (black arrow). (Photographs courtesy of Alexander M. Vaccaro, MD, Professor of Orthopaedic Surgery, Thomas Jefferson University and the Rothman Institute, Philadelphia.)
TREATMENT
Nonoperative Treatment
General Concepts
The crucial first treatment steps are timely injury recognition and determination of stability. Reduction maneuvers are typically performed with cranial skeletal traction in the emergency room using fluoroscopy. Traction is, however, contraindicated in distractive cervical spine injuries (77). Closed reduction of such distractive injuries may necessitate early application of a halo or postural reduction in a Rotorest bed or with sandbags surrounding the head, both of which are usually temporizing measures pending operative stabilization (19).
Accompanying resuscitation efforts include vasopressor support for suspected neurogenic shock and emergent assessment for potential intracranial trauma. Patients with neurologic injuries should be considered for intravenous methylprednisolone per the NASCIS III protocol, although the role of steroids in the treatment of acute spinal cord injuries has become increasingly unclear.
Emergent surgical intervention for patients with upper cervical spine injuries is rarely necessary. Open reduction and stable internal fixation are helpful intervention strategies for patients with dislocations and distractive upper cervical spine injuries. The presence of a spinal cord injury usually suggests the need for surgical stabilization and, possibly, decompression to maximize the chance for neurologic recovery.
Nonoperative treatment options consist of recumbent skeletal traction, bracing, and halo immobilization. The reduction can be assessed by obtaining lateral recumbent and upright x-rays (78). The duration of external immobilization usually ranges from 2 to 4 months, and depends on the type of injury and age of the patient. External immobilization is also commonly used for 6 to 12 weeks postoperatively after surgical stabilization. Recommendations vary widely with regard to the need for and duration of external support.
Bracing
In the presence of minimally or nondisplaced fractures of the upper cervical spine, external bracing alone can be considered (79). Several collars of varying shapes and rigidity as well as cervico-thoracic devices can be used, depending on the injury characteristics and patient-specific factors. SOMI-type devices have been shown to allow the least upper cervical spine motion of nonhalo devices in cadaveric testing (80).
Halo Orthosis
Halo-ring and vest orthotics offer the most stable form of external upper cervical spine immobilization (80). The halo has been recommended for patients with isolated occipital condyle fractures, unstable atlas ring fractures, odontoid fractures, and displaced neural arch fractures of the axis (81). Unlike bracing, the halo allows for some fracture manipulation and correction of malalignment. However, secondary loss of reduction has been noted in approximately one half of patients (78,82). A common mechanism of fracture displacement in a halo consists of a “snaking” of the cervical spine between supine and upright positions (83). Although this phenomenon may not adversely affect the healing of inherently stable upper cervical spine fractures with large cancellous bone surfaces (84), unstable fractures with a small bony contact surface, such as type II odontoid fractures, may not be effectively immobilized (12,77,85). Because the effectiveness of the halo-vest relies on a tight fit of the vest around the torso (86), it is poorly tolerated by elderly patients and patients with pulmonary compromise or thoracic deformities, such as with ankylosing spondylitis (82,87,88,89).
Skeletal Traction
Aside from its role in acute fracture reduction, cranial traction across either a halo-ring or Gardner-Wells tongs can be used to maintain spinal alignment and stability for extended periods to achieve initial consolidation of an unstable fracture prior to mobilizing the patient with a halo-vest or rigid brace (Fig. 38-20) (90). Although there are no fixed guidelines, suggested time frames for traction range from several days to weeks (75,82). However, prolonged recumbence carries an increased morbidity and mortality risk, and consideration should be given to the use of a Rotorest bed and mechanical as well as pharmacological thromboembolism prophylaxis (91). The role of prolonged traction has diminished progressively as stabilization techniques have become more versatile and comprehensive.
Indications for Surgical Treatment
Occipital Condyle Fractures
Operative treatment of occipital condyle fractures is generally reserved for the type III injuries that represent alar ligament avulsions which compromise craniocervical stability (Fig. 38-5). Surgical indications are therefore similar to those described below for craniocervical dissociation.
Craniocervical Dissociation
Displacement of greater than 2 mm at the atlanto-occipital joint, either on static imaging studies or with provocative traction testing (Fig. 38-6; Table 38-1), or the presence of neurologic injury are indications for craniocervical stabilization. Particularly in the presence of neurologic deficits, stabilization is performed as soon as safely possible in these often multiply injured patients (30).
Fractures of the Atlas
Most C1 fractures are treated by nonoperative methods. Operative management is usually necessary with loss of transverse atlantal ligament integrity (2,92), as suggested by a combined
P.1458

P.1459

lateral mass displacement of 7 mm or more (8.1 mm with standard x-ray magnification). Transverse atlantal ligament disruption introduces the potential for progressive lateral mass separation, C1–C2 instability, and pseudarthrosis (45,56,93). Halo immobilization alone may be insufficient to maintain acceptable alignment in these patients (94,95). If upright x-rays in a halo show further lateral mass displacement or an anterior atlanto-dens interval (ADI) of >3 mm, patients must be treated either with prolonged recumbency in cranial tong traction (Fig. 38-20) or with operative stabilization, generally with posterior C1–C2 or occiput-C2 fixation.
FIGURE 38-20 A 64-year-old man fell from a scaffold. He complained of neck pain. Physical examination revealed no neurologic deficits. A. The lateral x-ray shows obliquity of the atlas ring despite being well centered over the upper cervical spine (note the virtually superimposed pedicles of the axis). B. The open-mouth odontoid view confirms overhang of the C1 lateral masses by 11 mm relative to C2. C. The axial CT confirms the presence of a two-part atlas fracture with traumatic avulsion of the transverse atlantal ligament (TAL). Following a period of closed reduction with cranial traction, the definitive treatment was halo immobilization for 3 months. D. After halo removal, open-mouth odontoid x-ray shows persistent overhang of the lateral mass of the atlas. Postimmobilization dynamic extension (E) and flexion (F) x-rays demonstrate the atlantoaxial motion segment to be stable. The patient has noticed some neck stiffness but remains pain free.
Surgical stabilization options consist of C1–C2 transarticular screw fixation (96) or segmental fixation with C1 lateral mass and C2 pedicle screws connected by a plate or rod (97). The latter method provides the opportunity to correct the C1 lateral mass widening by approximating the two rods with a transversely oriented cross-connector. Internal fixation of the C1 ring, by simply reapproximating the lateral masses to each other through lateral mass screws connected to a transversely oriented rod (Fig. 38-21), is a potentially useful but not yet validated treatment option which theoretically preserves C1–C2 motion. A potential deficiency of this technique is that the associated TAL deficiency may result in persistent C1–C2 instability. However, unlike with shear or distractive injuries, the axial loading mechanism that causes transverse atlantal ligament rupture in displaced C1 ring fractures allows secondary restraints to remain intact, thus minimizing any remaining atlanto-axial instability once the atlas has been stabilized (98).
Atlanto-Axial Instability
Rotational instability is more commonly a result of infection rather than trauma. Traumatic rotational atlanto-axial injuries are usually associated with fractures such as type 2 odontoid fractures or a shear fracture of the inferior articular process of the atlas (Fig. 38-3). Fortunately, the critical craniocervical ligaments are usually intact in this injury.
Translational instability generally requires posterior atlanto-axial arthrodesis (Fig. 38-8). However, in the presence of bony avulsion at the TAL insertion onto the atlas, successful healing may occur in approximately three-fourths of patients with a period of recumbent traction followed by patient mobilization in a halo or SOMI (41). An ADI of >3 mm on flexion x-rays after 3 months of immobilization constitutes a failure of closed treatment and indicates the need for atlanto-axial arthrodesis.
Distraction injuries of C1–C2 with ≥l2 mm of displacement requires surgical stabilization. This injury is analogous to craniocervical dissociation at the atlanto-occipital joint and should be treated under similar guidelines. Chances for survival from this injury are poor due to common severe neurologic injury (Fig. 38-9).
Odontoid Fractures
Type I: Since the treatment of type I odontoid fractures relates to the effect of any associated alar ligament incompetence on craniocervical stability (68), indications for surgical management of these injuries are similar to those discussed for the treatment of craniocervical instability (Fig. 38-4).
P.1460

Type II: The management of type II odontoid fractures remains controversial. The preponderance of evidence suggests that surgical stabilization is appropriate for irreducible fractures (Fig. 38-22), fractures with distractive patterns of displacement (Fig. 38-23) (99,100) or fractures with associated spinal cord injury. Relative indications include multiply injured patients, associated closed head injury, initial displacement of 4 mm or more, angulation greater than 10 degrees (12,99,101,102), delayed presentation (>2 wk), multiple risk factors for nonunion (69,103,104,105), the inability to treat with a halo due to advanced age (50,91), associated cranial or thoraco-abdominal injury or other medical factors, and the presence of associated upper cervical fractures.
FIGURE 38-21 Open reduction and internal fixation of C1 lateral mass fracture. A. Axial CT image shows left lateral mass and posterior arch of C1 fractures in an elderly patient who could not tolerate rigid external immobilization. B. Postoperative axial CT image shows direct reapproximation and stabilization of the C1 fracture with C1 lateral mass screws connected by a transverse bar, as further illustrated with an atlas model (C). The indications for this procedure have not yet been well established.
Noncomminuted fractures in patients with favorable bone quality and appropriate body habitus are ideal for anterior odontoid screw fixation (106,107), which allows preservation of some atlanto-axial motion (Fig. 38-24) (108). In patients with extensive fracture comminution, compromised bone quality, or in whom achieving the requisite anterior odontoid screw trajectory is not feasible due to body habitus or the neck position required to maintain reduction, posterior atlanto-axial arthrodesis using either transarticular screw fixation or segmental C1–C2 fixation is indicated (97,109). Posterior atlanto-axial arthrodesis is the recommended treatment for type IIA dens fractures, which are inherently unstable due to a zone of segmental comminution at the odontoid
P.1461

P.1462

P.1463

base (69), and “sagittally oblique” fractures (110), in which the fracture line parallels the typical odontoid screwtrajectory, leading to loss of reduction and inadequate fixation with attempts at interfragmentary compression (Fig. 38-25) (111). Posterior C1–C2 arthrodesis with transarticular screws is expected to have the most predictably favorable result in the management of these two injury subtypes (83,109,110).
FIGURE 38-22 Type II odontoid fracture treated with posterior C1–C2 instrumented arthrodesis. This septuagenarian patient sustained a displaced type II odontoid fracture after a fall from standing. Acceptable fracture alignment could not be maintained with halo immobilization, as shown on lateral cervical spine x-ray (A) and sagittal CT image (B) after halo immobilization. C. Postoperative lateral x-ray after posterior C1–C2 instrumented arthrodesis. The presence of osteoporosis in this elderly patient was the major determining factor for selecting posterior atlantoaxial arthrodesis over odontoid screw fixation.
FIGURE 38-23 A. The cervical spine of this 350 lb man of short stature was cleared after a high-speed car crash based on interpretation of a standard series of cervical spine x-rays including this lateral image. Without cervical spine protection, the patient was transferred to a trauma center for treatment of an unreduced hip fracture dislocation. B. After arrival at the trauma center, repeat cervical spine x-rays showed a vertically distracted type II odontoid fracture. A solid C1–C2 fusion was demonstrated on open-mouth odontoid (C) and lateral (D) x-rays 1 year after atlantoaxial arthrodesis by the Brooks technique using four cables. The large body habitus of this patient precluded odontoid or transarticular screw instrumentation.
FIGURE 38-24 A. Lateral x-ray of an angulated and posteriorly displaced type II odontoid fracture in a 38-year-old man immobilized in a halo. After several failed attempts at closed reduction, anterior screw fixation was performed with the two-screw technique. As seen on these open mouth (B) and lateral (C) x-rays, the patient experienced uneventful healing of the fracture after immobilization in a rigid neck collar for 2 months.
Although one study has reported satisfactory results with odontoid screw fixation for fractures more than 3 months old (111), odontoid screw fixation beyond 3 weeks carries a substantially increased risk of nonunion, and thus is not a preferred treatment.
Proper patient selection helps minimize the high complication rate of up to 28% that has been reported using anterior odontoid screw fixation (106,110,112,113).
Type III: Type III odontoid fractures rarely require surgical stabilization. As with type II odontoid fractures, operative stabilization is indicated in fractures with associated spinal cord injury or distractive instability patterns (Fig. 38-26) (114). Posterior C1–C2 arthrodesis is the surgical treatment method of choice, since anterior odontoid screw fixation has a high failure rate with these injury types (111). Relative indications include highly displaced irreducible fractures, patients with displaced injuries who cannot be treated with a halo for reasons cited earlier, and fractures with initial
P.1464

P.1465

P.1466

displacement of 5 mm or more, which have a high potential for nonunion, particularly in the elderly population (101). Results may be less predictable than is generally recognized, however, with delayed unions or pseudarthroses reported in up to 54% of nonoperatively treated patients (101).
FIGURE 38-25 Anterior screw placement for odontoid fractures. A. The starting point (large white arrow) is posterior to the anterior cortical lip of the axis. The correct angulation of the screw trajectory is tilted only slightly posterior to the longitudinal axis of the odontoid. Care should be taken to achieve fracture compression by having the screw threads engage bone only past the fracture. The cortex of the odontoid tip should be engaged or barely penetrated by the screw thread. Different screw placement techniques are necessary for the single-screw technique (gray arrow) compared to the two-screw technique (black lines). B,C. While the single-screw technique aims to have a screw placed in the center of the midaxial line, the two-screw technique usually necessitates a slight convergence of the two screw tips due to the narrowness of the odontoid waist. Careful patient selection is an important prerequisite for successful outcomes with this technique. Among many other potential factors an anterior oblique fracture (C) can lead to compromised fixation stability.D. Exacting surgical technique and intraoperative radiographic visualization are facilitated by biplanar imagingas provided by two C-arms. The feasibility of anterior odontoid screw fixation in patients with a prominentchest or with pronounced cervical kyphosis can be evaluated by preoperatively placing a radio-opaque markeralong the fluoroscopically confirmed intended screw trajectory.
FIGURE 38-26 Type III odontoid fracture with craniocervical distraction. Sagittal CT (A) and MRI (B) images show a type III odontoid fracture with 3 mm of distraction at the fracture site (white arrow). This atlantoaxial distractive injury is associated with extensive ligamentous disruption, as illustrated by the posterior element widening and increased signal intensity between C1–C2 on MRI (gray arrow). C. A lateral x-ray 6 months after posterior instrumented atlantoaxial arthrodesis shows that anatomic cranio-cervical alignment has been maintained.
Traumatic Spondylolisthesis of the Axis(Hangman’s Fractures)
Operative stabilization is rarely indicated for traumatic spondylolisthesis of the axis (76,115). Most injuries can be treated with early ambulatory immobilization (115) with 12 weeks of bracing for type I and most type IA fractures, and a similar period of halo immobilization for most type II fractures (73). Type IIA injuries may be treated with halo immobilization if their alignment can be successfully maintained (Fig. 38-27). Traction is contraindicated in type IIA injuries, as it accentuates their kyphotic deformity.
If the kyphotic deformity in type IIA injuries cannot be controlled in a halo, surgical intervention is indicated. A C2–C3 anterior discectomy and arthrodesis (ACDF) with plating allows for arthrodesis across the least number of levels, and preserves atlanto-axial motion (116,117,118). However, because the anterior longitudinal ligament is often the only remaining intact C2–C3 stabilizing structure, posterior stabilization remains an appropriate option. A disadvantage of the posterior approach is that, absent the ability to gain acceptable purchase with C2 screws directly across the fracture, the need to extend fixation to the C1 level results in loss of atlanto-axial motion (Fig. 38-17).
Type III injuries usually cannot be reduced by traction and require operative reduction and stabilization. Stabilization options include (i) posterior C1–C3 arthrodesis (Fig. 38-28); (ii) posterior C2–C3 arthrodesis using interfragmentary C2 screws across the pars interarticularis fractures; (iii) converting the fracture to a type I or II injury by posterior C2–C3 facet arthrodesis using C2 screws which stop short of the fracture, followed by collar or halo immobilization per the usual treatment for type I and II injuries; and (iv) anterior C2–C3 ACDF (Fig. 38-18), in the unusual event that reduction occurs by closed methods. The advantage of the latter three options is that they preserve atlanto-axial motion (75,119).
P.1467

P.1468

P.1469

P.1470

P.1471

P.1472

P.1473

P.1474

P.1475

P.1476

P.1477

P.1478

P.1479

P.1480

P.1481

P.1482

P.1483

P.1484

P.1485

TREATMENT RESULTS AND COMPLICATIONS
The clinical outcome of patients who have sustained upper cervical injuries is often more dependent on associated intracranial injury than on their spinal injury. Additional patient comorbidities add to the confusion with regard to realistically attainable clinical results. An understanding of specific problems associated with specific injury types is, however, helpful in preventing their occurrence.
Injury-Specific Treatment Results and Complications
Occipital Condyle Fractures
Little has been written about the morbidity associated with occipital condyle fractures (151,152). Outcome is likely to depend on the potential presence of various comorbidities, such as associated head injury. An association between occipital condyle fractures and palsy of their most closely associated cranial nerves (IX, X, XI, XII) in up to a third of cases has also been described (1,8,153,154). Although spine-related patient outcome is presumably contingent on the presence or absence of symptoms of post-traumatic arthritis such as neck pain, occipital headaches, and restriction in craniocervical motion, the incidence of posttraumatic arthritis is unknown (155). Torticollis may result from chronic atlanto-occipital subluxation.
The outcome of type III fractures that are components of distractive injuries is better represented in the subsequent discussion of craniocervical dissociation outcomes.
Craniocervical Dissociation
The majority of craniocervical dissociations are fatal. The outcome of survivors depends on (i) the type and severity of associated injuries, particularly closed head injuries, (ii) the severity of spinal cord injury, and (iii) the timely diagnosis and treatment of craniocervical dissociation (30). Since complete neurologic injury at the craniocervical junction is likely to result in fatal cardiorespiratory arrest, neurologic deficits and the degree of craniocervical displacement are likely to be less severe in survivors. Partly because of these reasons and despite substantial advances in neuroimaging over the last two decades, delayed diagnosis of craniocervical dissociation continues to be a frequent
P.1486

problem (30,156,157). Early recognition and timely fixation of these injuries, both of which may be elusive in these severely polytraumatized patients, improves outcome by protecting against neurologic deterioration (30). A delayed diagnosis in these highly unstable injuries has been associated with secondary neurologic deterioration and possibly death in up to 75% of patients (30,158,159). These unacceptably high numbers underscore the importance on improving our current cervical spine trauma screening measures. Our clinical awareness of injuries to the cervical spine should fully incorporate the understanding of injuries of the craniocervical junction.
Fractures of the Atlas
Severe complications associated with an isolated fracture of the atlas are fortunately rare. With the exception of pain or loss of sensation in the greater occipital nerve distribution, neurologic sequelae are uncommon and more likely to be related to associated injuries (160). Patients with posterior arch fractures are expected to heal with few or no symptoms. Osteoarthritis of the atlanto-occipital and occipito-cervical joints may occur as a result of a displaced lateral mass fracture. Minimally displaced lateral mass or Jefferson (burst) fractures treated nonoperatively are associated with an 80% incidence of residual neck pain, and a 17% nonunion rate (63). The amount of displacement that can be tolerated by the upper cervical spine articulations without developing posttraumatic arthritis is unknown. Severe malunion of an unstable atlas fracture resulting in a painful torticollis, such as with a highly displaced lateral mass fracture, may require realignment and occiput to C2 arthrodesis (Fig. 38-40).
Atlanto-Axial Instability
Traumatic Transverse Ligament Insufficiency Acute traumatic rupture of the TAL caused by translational forces, in contrast to that which occurs due to axial loading with displaced fracture of the atlas, is usually fatal (161). In the occasional survivor, profound neurologic deficits may be present (162). Because of the common mechanism of a blow to the occiput causing forced forward flexion and translation of C1 on C2, associated head injuries are common and may be the predominant influence on outcome in survivors (98,161). Syncope and vertigo may result from injury to the vertebrobasilar arterial system.
Patients with type II bony avulsion injuries of the TAL have a 26% reported likelihood of instability after nonoperative treatment (41). Atlanto-axial arthrodesis can be considered in patients with dynamic x-rays indicating late atlanto-axial instability or in patients with painful atlanto-axial arthritis confirmed on CT scan and bone scan. In the case of long-standing and irreducible deformities, adjunctive decompression and cranial extension of the arthrodesis to the occiput may be required.
TAL rupture may occur in a minority of odontoid fractures. This combination of injuries has been described as an indication for early C1–C2 stabilization due to the high likelihood of persistent C1–C2 instability even with a successfully healed odontoid process (40). Atlanto-Axial Distraction The prognosis for outcome and complications with this injury pattern are best represented by the section on craniocervical dissociation outcomes.
Odontoid Fractures
Odontoid fractures continue to be associated with significant morbidity and even mortality. Fracture nonunion and missed injuries are the most common causes of complications (163). Primary neurologic injury or secondary deterioration are rarely encountered (130,164). Pseudarthrosis of a type II odontoid fracture is a leading cause of secondary neurologic deterioration (130). A pseudarthrosis of the odontoid has been defined as the absence of fracture site bridging after 4 months of treatment (165). Most cases of os odontoideum may in fact represent a nonunion of a type I or II odontoid fracture (68,166).
Type I injuries are rarely encountered. Based on the limited number of cases described in larger studies on odontoid fractures (12,167) in the absence of associated craniocervical instability, few complications or long-standing symptoms result from treatment of an isolated type I fracture with external immobilization.
Type II odontoid fractures have been associated with a high nonunion rate regardless of the type of nonoperative treatment. Without immobilization, Type II fractures have been found to have a 100% risk of nonunion (12,167). Nonoperative treatment consisting of either bracing or halo has been reported to result in nonunion rates ranging from 15% to 85% (12,68,165). Prerequisites for a successful nonoperative treatment are maintenance of a nearly anatomic fracture reduction without distraction (12,168). More than 20% translation of a type II odontoid fracture provides insufficient fracture surface contact for bony union (169). Of many potential predisposing risk factors for fracture nonunion, displacement of 4 to 5 mm has been the most consistently identified factor (12,165,167). Other risk factors for nonunion include age above 60 years, fracture angulation greater than 9 degrees, posterior fracture displacement, and delay in treatment (83,110,168,169,170,171).
Failure of fracture healing with odontoid screws has been described in 10% of patients, with an overall perioperative complication rate of up to 28% (110,113). C1–C2 arthrodeses have reported nonunion rates of 4% or less using transarticular screw and wired structural bone-graft constructs (12,109,172).
Nonoperative treatment of type III odontoid fractures in a halo is associated with pseudarthrosis rates from 9% to 13% (12,68). However, fracture displacement ≥l4 mm or angulation ≥l10 degrees have been associated with nonunion rates of 22% to 54% (12,101,165). If surgical stabilization of type III odontoid fracture is undertaken, it should be in the form of atlanto-axial fixation, since excessively high failure rates (55%) have been reported for odontoid screw fixation (111).
Although spinal cord injury as a result of fractures in the upper cervical spine is less common than in the lower cervical or thoracic spine (164), type II odontoid fractures are the most common nondistractive fractures of the upper cervical spine to present with primary neurologic injury. The frequency of
P.1487

P.1488

neurologic injury with type II odontoid fractures ranges from 18% to 25% (12,68), and the severity of ranges from isolated cranial nerve injury to pentaplegia.
FIGURE 38-40 Nonunion of C1 lateral mass fracture with resulting craniocervical malalignment. A. Anteroposterior x-ray of a patient who sustained a right C1 lateral mass fracture that was treated with a halo-vest. She developed progressive tilt and rotation at the craniocervical junction and occipital dysesthesias. B. Coronal CT image and (C) axial CT image illustrate the malalignment of the craniocervical functional unit, with lateral displacement of the right C1 lateral mass, which results in articulation of the right occipital condyle directly with the C2 lateral mass. The patient underwent corrective osteotomy with instrumented posterior craniocervical arthrodesis, as illustrated on anteroposterior (C) and lateral (D) x-rays 1 year postoperatively.
Considerable mortality is associated with upper cervical spine fractures. Despite the increased likelihood of fatality in healthy individuals sustaining high-energy injuries, the high reported mortality rates appear to be primarily related to the high proportion of elderly patients in whom these fractures may occur as a marker for progressively frail health. The in-hospital mortality rate for elderly patients with type II odontoid fractures has been reported to be 27% to 42% (91,168,173,174). In one series, however, this high mortality rate was reduced to zero after institution of an early surgical stabilization protocol (91). Early mobilization with a neck collar, with surgical stabilization when appropriate, is considered the preferred treatment strategy for upper cervical spine injuries in this age group (91,174).
Traumatic Spondylolisthesis of the Axis (Hangman’s Fracture)
Although fractures of the axis account for 25% to 71% of deaths at the scene of injury (2,3,167) acute post-admission mortality after hangman’s fractures is thought to be as low as 2% to 3% (167). Neurologic injury resulting from this fracture type has been identified in 3% to 10% of patients (51,76). Type III injuries, which have a 60% reported rate of associated spinal cord injury, are at highest risk for neurologic deficit due to the facet dislocation component (75). Type IA fractures, however, reportedly have a 33% incidence of associated spinal cord injury, probably due to the canal compromise which occurs with fracture displacement in this atypical oblique fracture pattern (Fig. 38-19). Type IA injuries also have a greater potential for vertebral artery injury because the foramen transversarium is commonly involved on one side (74).
Successful healing of C2 traumatic spondylolisthesis is reported to approach 95% (51). This is most commonly achieved with nonoperative measures. Associated upper cervical (15%), subaxial (23%), or head injuries are usually greater contributors to prognosis than the C2 fracture itself. Fractures of types IA, IIA, and III subcategories constitute a greater treatment challenge due to either their atypical fracture orientation or their inherent ligamentous injury component.
Symptomatic degeneration of the C2–C3 articulations, potentially requiring arthrodesis, is thought to occur in 10% of patients, and is more likely to occur in type I injuries since most patients with type II injuries progress to spontaneous anterior ankylosis (69). This may explain why, in patients with displaced injuries, long-term symptoms are rare despite the frequent absence of bony healing across the pars fracture. Patients who heal in severe kyphosis may have difficulty with neck extension.
Symptomatic pseudarthrosis is unusual. In the case of type I injuries, treatment involves either direct osteosynthesis with interfragmentary compression across the pars interarticularis fracture through a posterior approach, or C2–C3 arthrodesis through an anterior approach (76). In type II injuries, because of the greater deformity and displacement and the associated injury to the C2–C3 intervertebral disc, anterior arthrodesis is usually more appropriate than attempts at direct osteosynthesis.
Combined Injuries
Combined craniocervical injury patterns occur with relative frequency (Fig. 38-41). Their treatment outcomes depend on the specific characteristics of each individual injury, but more importantly, on the timely recognition of these complex injury patterns, which allows for their appropriate treatment.
Vascular Injuries
Vascular injuries are not infrequent with upper cervical spine trauma, although the incidence remains unclear and depends on the diagnostic modalities used (54,175). Vertebral artery disruption should be specifically considered in any distractive upper cervical spine injury, such as atlanto-axial dissociation or in patients with type III or IV atlanto-axial rotary subluxation. In addition, vertebral artery injuries should be considered in any displaced fracture involving the transverse foramen. Despite the vertebral artery running in close proximity to the rostral lamina of the posterior ring of the atlas, fractures to the atlas ring have not been commonly associated with local vertebral artery trauma. Forced hyperflexion injuries, such as with anteriorly displaced type III odontoid fractures, may leave the vertebral arteries spared, but can lead to thrombosis of the carotid arteries (176).
Treatment-Specific Results and Complications
Skeletal Traction and Halo-Vest
Complications associated with skeletal traction include local complications or systemic ailments resulting from prolonged recumbency. Local complications include pin tract infection, loss of fixation, occipital decubitus, dural pin penetration, and propagation of a skull fracture. Systemic complications associated with prolonged skeletal traction include respiratory compromise, thromboembolism, decubitus ulcers, and sepsis (88,89). Pharmacological thromboembolism prophylaxis in patients with acute spine fracture is associated with an increased risk of epidural hematoma formation. In general, prolonged traction for upper cervical spine fractures has become increasingly unpopular, despite the absence of specific data delineating the incidence of medical complications. Judicious use of cervical traction therefore remains a valid treatment option under certain circumstances. For elderly patients prolonged traction is associated with a significant morbidity and is usually less desirable (91).
Halo-vests have relatively high associated complications. These include loss of reduction in 46% of patients, pressure sores in 11% (Fig. 38-42), and dural pin penetration in 1% of patients. Pin complications are commonly encountered: loosening has been noted in 36% of patients, infections in 20%, pain at the pin site in 18%, and disfiguring scars in 9%. Increased incidence of pulmonary complications and aspiration are associated with use of halo-devices in elderly patients (91).
P.1489

P.1490

P.1491

Craniocervical Arthrodesis
Craniocervical dissociation is fatal in the majority of cases, allowing for few meaningful descriptions of treatment results. Pseudarthrosis rates of up to 23% have been reported with onlay grafting and nonrigid fixation (wiring) methods (177,178,179,180,182). An 89% fusion rate has been reported using onlay grafting alone, which eliminates hardware-associated complications but requires the use of aggressive postoperative immobilization techniques including recumbency and skull tong traction, Minerva jacket, and halo (183,184). Onlay structural autograft with cerclage wire alone resulted in comparatively improved fusion rates (181), but had the disadvantage of requiring more comprehensive postoperative external immobilization, and was complicated by wire breakage in 78% of patients (185) and late fracture of the graft in up to 15% of patients (186). These problems were initially addressed by substituting a contoured loop for onlay bone graft (187,188).
FIGURE 38-41 Multiple craniocervical injury patterns may occur concurrently, and must be sought out by the treating clinician. An adult female was involved in a high-speed motor vehicle collision, and the diagnosis of type III odontoid fracture was made. A. Lateral cervical spine x-ray, (B) sagittal CT reformation, and (C) coronal CT reformation show the odontoid fracture (gray arrow) as well as distraction of the atlanto-axial (white arrows) and atlanto-occipital (black arrows) joints. The associated craniocervical distractive injury was not initially recognized. The patient was treated provisionally in a halo-vest and remained neurologically intact. When the initially unrecognized extent of her craniocervical dissociation was appreciated, as confirmed by distraction and increased signal intensity at the atlanto-occipital and atlantoaxial joints on (E) sagittal and (F) coronal MRI scans, she underwent occipitocervical stabilization. Postoperative sagittal midline (G) and parasagittal (H) CT reformations show reduction and stabilization of the craniocervical junction, which has been maintained on lateral cervical x-rays (I) 6 months postoperatively.
FIGURE 38-42 Extensive scarring around halo pin tract site. This patient was treated with a halo for a type III odontoid fracture. The anterior halo pins were erroneously placed in the temporal fossae. Extensive ulceration was noted around these anterior pins bilaterally. The right side is shown (black arrow) after 3 months of local wound care.
The use of rigid fixation with plates or rods and screws has improved nonunion rates to less than 6% (180,182). More recent case series have shown that rigid craniocervical arthrodesis techniques comprised of screw and plate constructs with adjunctive suboccipital- and sublaminar-cabled structural graft has resulted in fusion rates approaching 100% with no incidence of hardware failure or need for revision surgery due to reasons of instability despite, in general, the inclusion of fewer motion segments (30,189,190). Potential technical problems include malreduction, which may result in neurologic worsening (30) and possible penetration of the inner cortex of the skull which can lead to injury to neural or vascular structures. Recent series suggest, however, that the bigger challenge in treating survivors of craniocervical dissociation lies not in their surgical treatment per se, but in recognizing their often radiographically subtle yet highly unstable injuries and maintaining sufficient stability to protect neurologic function during the initial pre-operative treatment phase, which almost always includes the resuscitation and multi-system evaluation in these polytrauma-tized patients (30). Posterior craniocervical stabilization has recently been shown to be neuroprotective after craniocervical dissociation as compared to an approximately 40% incidence of neurologic worsening in patients who had a delay in diagnosis.
Posterior Atlanto-Axial Arthrodesis
Structural Bone Graft and Wiring Wiring techniques are a safe and straightforward method of atlanto-axial stabilization. However, pseudarthrosis occurs in up to 25% of cases despite the use of more rigid adjunctive external immobilization methods (134,136,191). This problem is related to the relative inability of posterior wiring alone to neutralize mainly rotational, and to some extent translational forces (165,192,193,194). Another described complication is extension of the fusion mass to the occiput (195).
Posterior wiring techniques are not effective for injuries where the atlas is displaced posteriorly relative to the axis. In such injuries, posterior wiring tends to accentuate rather than correct the deformity.
Loss of reduction has also been reported to occur with the use of posterior bone-graft and wiring techniques alone (193). Although uncommon, spinal cord injury may occur, presumably during wire passage, and may be avoided by not passing wires before complete reduction has been achieved (12). Transarticular Screws The advent of transarticular screw fixation allowed a greater degree of atlanto-axial stability than was previously achievable (131,145,192,196). Fusion rates approaching 100% have been reported following transarticular screw fixation, usually with adjunctive posterior bone grafting and wiring techniques (109,172). The main concerns pertain to the potential for screw malposition resulting in injury to the vertebral artery, spinal cord, hypoglossal nerve, and carotid artery.
The proximity of the vertebral artery to the transarticular screw pathway has led to a reported incidence of vertebral artery injury as high as 6%, with an associated 0.2% risk of immediate neurologic deficit and 0.1% mortality (123,197,198). Should a vertebral artery injury be suspected, local hemostasis can be achieved with bone wax or placement of the transarticular screw
P.1492

on the affected side. There should, however, be no attempt to place the contralateral transarticular screw. Postoperatively, selective angiographic embolization can assure satisfactory hemostasis and prevent occurrence of an arterio-venous fistula and embolic stroke.
Excessive drill penetration of the anterior cortex of the atlas can lead to injury of the internal carotid artery or the hypoglossal nerve (109). Carotid artery injury is not likely to be recognized intraoperatively. Excessive retropharyngeal soft tissue swelling as seen on a lateral cervical x-ray can lead to the diagnosis. Angiographically aided embolization following trial occlusion with an intra-arterial balloon can be used to control local hemorrhage. These complications can be avoided by appreciating that the anterior cortex of the C1 lateral mass lies posterior to the anterior most projection of the C1 anterior arch on a lateral x-ray by an average of 7 mm. This distance can be estimated by evaluating preoperative axial images of the atlas (143).
Fusion rates with transarticular screw fixation are reported to be above 95%. Early hardware failure is an unusual occurrence which usually consists of cutout of the screw tip in the atlas or of the screw shaft in the pars of the axis. Anatomic patient variations such as a very shallow and arcuate isthmus can predispose to limited screw purchase within the axis. Similarly an insufficiently cephalad drill trajectory can lead to limited screw purchase at the antero-inferior edge of the C1 lateral mass. Late hardware failure can result from a pseudarthrosis, the salvage of which may consist of revision posterior atlanto-axial arthrodesis, occipito-cervical arthrodesis, or anterior atlanto-axial arthrodesis (124,127,199,200). Removal of a broken distal transarticular screw is usually unnecessary and impractical.
Proximity of the C2 nerve root to the posterior isthmus of the atlas exposes it to potential injury during posterior C1–C2 facet arthrodesis. The resulting symptoms of occipital numbness or dysesthesias, however, have only rarely been reported (109,172).
Approximately 50% loss of head rotation can be expected following successful C1–C2 arthrodesis (201). Although this may lead to functional impairment, the consequences have not been well-documented in the posttraumatic patient population.
C1 Lateral Mass and C2 Pedicle Screws Fixation which connects individual screws in the C1 lateral masses to C2 pedicle, or pars, interarticularis screws has become more popular over the last decade, presumably because of the greater versatility this technique affords as compared to transarticular screw fixation (132). This construct alone has been noted to be biomechanically similar to transarticular screws with adjunctive posterior bone graft and wiring (146). The main advantages of this technique relate to the use of a pedicle screw at C2, which has a more easily achievable trajectory and carries a lower theoretical risk of vertebral artery injury than the transarticular screw. Although this is a useful primary method of atlanto-axial fixation, this stabilization method is particularly valuable as an alternative to transarticular screw fixation when the latter procedure is deemed unsafe or technically implausible due to anomalous vertebral artery anatomy or when issues related to patient anatomy or positioning interfere with acceptable transarticular screw trajectory. Only small, uncontrolled case series have been published to date using this technique, with equivalently high fusion and low loss of fixation rates to transarticular screw techniques (97,132,147).
Anterior Approach
Risks of the anterior surgical approach to the neck include neurologic injury, and injury to the esophagus, trachea, or vascular structures. All of these are relatively infrequent events and should occur in less than 5% of patients. Dysphagia as a result of anterior surgical exposure is relatively common, and is more prevalent in the upper than in the lower cervical spine (126,127,150,202). There does not seem to be any difference in the frequency of recurrent laryngeal nerve injury between right- vs left-sided upper cervical spine approaches.
Anterior Odontoid Screw Fixation
Results of anterior odontoid screw fixation have varied in the literature. Although some large series have shown excellent success rates, with healing rates approaching 90% in patients treated within 6 months of injury regardless of patient age and bone quality (111,203,204), other series have shown higher complication rates, particularly loss of fixation, in patients with osteoporosis (205). In addition to bone quality, sagittal plane fracture obliquity in the anterior direction similar to that of intended screw placement appears to be associated with lower (75%) healing rates (111). In general, patients with anteriorly displaced odontoid fractures pose a greater challenge to odontoid screw fixation, compared to those with posteriorly displaced fractures. Lack of an anatomic reduction or inability to achieve interfragmentary compression across the fracture can greatly impair the efficacy of odontoid screw fixation.
Major complication rates of up to 28% have been reported with odontoid screw fixation (106,110,111,112,113,121,163) and consisted mainly of hardware-related complications (10%) and superficial wound infections (2%) in the largest published series (111). Of the hardware-related complications, one-half consisted of screw disengagement from the C2 body in patients with type III odontoid fractures, putting into question whether this technique is appropriate for odontoid fractures that extend into the C2 body. A second commonly seen hardware-related complication is loosening of the odontoid screw, which occurs primarily when the screw tip has not engaged the apical cortex of the odontoid. Failure of screw fixation may have catastrophic consequences, as demonstrated by a report of quadriplegia and death from respiratory failure that resulted from fracture displacement after loss of fixation (111). Similar sequelae of odontoid screw fixation have been reported on other occasions (203). A third hardware-related complication pertains to screw malposition. The failure rate of odontoid screw fixation does not appear to be influenced by the use of one versus two screws (111,121,206). Other technical mistakes include an excessively anterior starting point, which can leave a thin anterior bony
P.1493

shell in the axis that is unable to contain the screw shaft or require a screw trajectory that is too posteriorly oriented. If a cannulated screw system is used, care should be taken to avoid advancement of a guide-wire rostral to the tip of the odontoid. Although rare, intraoperative spinal cord and cranial nerve injury have been described (110,112,113). It is apparent from the literature that a considerable learning curve exists with this complex procedure (12,110).
Although fracture union does not appear to be affected by delay in surgery of up to 6 months, long-standing odontoid pseudarthrosis responds poorly to anterior screw fixation, as demonstrated by a mere 25% healing rate in a series of 18 patients who were operated for pseudarthrosis between 18 and 48 months post-injury (111). The 25% hardware-related complication rate in this group of patients, consisting mainly of screw fracture, was consistent with previously reported experiences with odontoid screw fixation of pseudarthroses (110).
Complications already described for the Smith-Robinson approach, such as esophageal or neurovascular injury, dysphagia, and pharyngeal edema have been reported to occur when placing anterior odontoid screws (110,111,203,207).
Technical considerations that may preclude the use of odontoid screw fixation pertain primarily to physical characteristics that do not allow for acceptable screw trajectory, such as prominent thoracic kyphosis, barrel chest, and fracture characteristics that require a flexed position to maintain an acceptable reduction.
C2–C3 Anterior Discectomy and Arthrodesis
As surgery is not commonly necessary for the treatment of traumatic spondylolisthesis of the axis, there are few reports on the results of anterior C2–C3 discectomy and arthrodesis for the treatment of traumatic conditions. One series of five patients (118) describes the successful use of this technique without complications, using autologous tricortical iliac crest graft and anterior plating with subsequent external immobilization in a halo, for the treatment of type II traumatic spondylolisthesis of the axis and associated upper cervical injuries. Another broader series describing anterior arthrodesis and osteosynthesis for cervical spine trauma included seven patients treated with C2–C3 ACDF and plating; five for traumatic spondylolisthesis of the axis and two for C2–C3 subluxation or dislocation. Six patients healed without complication whereas one patient died for reasons unrelated to the procedure. Combining these two series, all three patients presenting with neurologic deficits had profound postoperative recovery.
Complications related to this procedure pertain to the anterior approach to the upper cervical spine, which are well described (150), and to the potential for pseudarthrosis or loss of fixation, the frequency of which is not well documented. Upper cervical approaches appear to have a higher risk of airway-related problems (150). Due to small patient numbers there is also no comparison to posterior instrumentation approaches for the same conditions. The mandible commonly obstructs the surgical exposure of the anterior upper neck, adding to the complexity of the procedure. Reported problems with anterior C2–C3 arthrodesis are primarily technical, such as achieving an adequate anterior decompression, proper bone graft positioning, and placement of stable low-profile instrumentation. Prominent hardware could potentially lead to swallowing difficulties and even esophageal erosion, and should therefore be avoided. Horner’s syndrome and suboccipital pain with associated C2–C3 degenerative changes have been reported as frequent complications of this procedure for the treatment of traumatic spondylolisthesis of the axis (116).
REFERENCES
1. Ahuja A, Glasauer FE, Alker GJ Jr, et al. Radiology in survivors of traumatic atlanto-occipital dislocation. Surg Neurol 1994;41:112–118.
2. Alker GJ, Oh YS, Leslie EV, et al. Postmortem radiology of head neck injuries in fatal traffic accidents. Radiology 1975;114:611–617.
3. Bucholz RW, Burkhead WZ. The pathological anatomy of fatal atlanto-occipital dislocations. J Bone Joint Surg Am 1979;61:248–250.
4. Dvorak J, Schneider E, Saldinger P, et al. Biomechanics of the craniocervical region: The alar and transverse ligaments. J Orthop Res 1988;6:452–461.
5. Panjabi M, Dvorak J, Crisco J 3rd, et al. Flexion, extension, and lateral bending of the upper cervical spine in response to alar ligament transections. J Spinal Disord 1991;4:157–167.
6. Panjabi M, Dvorak J, Crisco JJ 3rd, et al. Effects of alar ligament transection on upper cervical spine rotation. J Orthop Res 1991;9:584–593.
7. Panjabi MM, Crisco JJ 3rd, Lydon C, et al. The mechanical properties of human alar and transverse ligaments at slow and fast extension rates. Clin Biomech (Bristol, Avon) 1998;13:112–120.
8. Crisco JJ 3rd, Panjabi MM, Dvorak J. A model of the alar ligaments of the upper cervical spine in axial rotation. J Biomech 1991;24:607–614.
9. Panjabi MM, Oxland TR, Parks EH. Quantitative anatomy of cervical spine ligaments. Part I. Upper cervical spine. J Spinal Disord 1991;4:270–276.
10. Alker GJ Jr, Oh YS, Leslie EV. High cervical spine and craniocervical junction injuries in fatal traffic accidents: A radiological study. Orthop Clin North Am 1978;9:1003–1010.
11. Beckner MA, Heggeness MH, Doherty BJ. A biomechanical study of jefferson fractures. Spine 1998;23:1832–1836.
12. Clark CR, White AA 3rd. Fractures of the dens. A multicenter study. J Bone Joint Surg Am 1985;67:1340–1348.
13. Dvorak J, Panjabi MM, Hayek J. [Diagnosis of hyper- and hypomotility of the upper cervical spine using functional computerized tomography]. Orthopade 1987;16:13–19.
14. Maynard FM Jr, Bracken MB, Creasey G, et al. International standards for neurological and functional classification of spinal cord injury. American spinal injury association. Spinal Cord 1997;35:266–274.
15. Brodke DD, Albert T. Upper cervical spine fractures in patients with spinal cord injury. Spine: State of the Art Reviews 1999;13:70–83.
16. Blacksin MF, Lee HJ. Frequency and significance of fractures of the upper cervical spine detected by CT in patients with severe neck trauma. AJR Am J Roentgenol 1995;165:1201–1204.
17. Harris JH, Carson GC, Wagner LK. Radiologic diagnosis of traumatic occipitovertebral dissociation: 1. Normal occipitovertebral relationships on lateral x-rays of supine subjects. AJR Am J Roentgenol 1994;162:881–886.
18. Monu J, Bohrer SP, Howard G. Some upper cervical spine norms. Spine 1987;12:515–519.
19. Chapman JR, Newell DN. Emergent management of spinal cord injury. Spine: State of the Art Reviews 1999;13:50–75.
20. Kirshenbaum KJ, Nadimpalli SR, Fantus R, et al. Unsuspected upper cervical spine fractures associated with significant head trauma: Role of CT. J Emerg Med 1990;8:183–198.
21. Deliganis AV, Mann FA, Grady MS. Rapid diagnosis and treatment of a traumatic atlantooccipital dissociation. AJR Am J Roentgenol 1998;171:986.
22. Deliganis AV, Baxter AB, Hanson JA, et al. Radiologic spectrum of craniocervical distraction injuries. Radiographics 2000;20 Spec No:S237–250.
23. Powers B, Miller MD, Kramer RS, et al. Traumatic anterior atlanto-occipital dislocation. Neurosurgery 1979;4:12–17.
24. Lee C, Woodring JH, Goldstein SJ, et al. Evaluation of traumatic atlantooccipital dislocations. AJNR Am J Neuroradiol 1987;8:19–26.
25. Harris JH Jr, Carson GC, Wagner LK, et al. Radiologic diagnosis of traumatic occipito-vertebral dissociation: 2. Comparison of three methods of detecting occipitovertebral relationships on lateral radiographs of supine subjects. AJR Am J Roentgenol 1994;162:887–892.
26. Dickman CA, Hadley MN, Pappas CT, et al. Cruciate paralysis: A clinical and radiographic analysis of injuries to the cervicomedullary junction. J Neurosurg 1990;73:850–858.
27. Davis JW, Parks SN, Detlefs CL, et al. Clearing the cervical spine in obtunded patients: The use of dynamic fluoroscopy. J Trauma 1995;39:435–438.
P.1494

28. Davis JW, Kaups KL, Cunningham MA, et al. Routine evaluation of the cervical spine in head-injured patients with dynamic fluoroscopy: A reappraisal. J Trauma 2001;50:1044–1047.
29. Harris MB, Kronlage SC, Carboni PA, et al. Evaluation of the cervical spine in the polytrauma patient. Spine 2000;25:2884–2891; discussion 2892.
30. Bellabarba C, Mirza SK, Mann FA, et al. Survival after craniocervical dissociation. 70th Annual Meeting of the American Academy of Orthopaedic Surgeons. New Orleans, LA: AAOS; 2003.
31. Chapman JR, Bellabarba C, Newell DW, et al. Craniocervical injuries: Atlanto-occipital dissociation and occipital condyle fractures. Sem Spine Surg 2001;13:90–105.
32. Mace SE. Unstable occult cervical-spine fracture. Ann Emerg Med 1991;20:1373–1375.
33. Mace SE. The unstable occult cervical spine fracture: A review. Am J Emerg Med 1992;10:136–142.
34. Panjabi MM, Dvorak J, Crisco J, et al. [Instability in injury of the alar ligament. A biomechanical model]. Orthopade 1991;20:112–120.
35. Ehara S, el-Khoury GY, Clark CR. Radiologic evaluation of dens fracture. Role of plain radiography and tomography. Spine 1992;17:475–479.
36. Wasserberg J, Bartlett RJ. Occipital condyle fractures diagnosed by high-definition CT and coronal reconstructions. Neuroradiology 1995;37:370–373.
37. Przybylski GJ, Clyde BL, Fitz CR. Craniocervical junction subarachnoid hemorrhage associated with atlanto-occipital dislocation. Spine 1996;21:1761–1768.
38. Kathol MH. Cervical spine trauma. What is new? Radiol Clin North Am 1997;35:507–532.
39. Katzberg RW, Benedetti PF, Drake CM, et al. Acute cervical spine injuries: Prospective MR imaging assessment at a level 1 trauma center. Radiology 1999;213:203–212.
40. Greene KA, Dickman CA, Marciano FF, et al. Transverse atlantal ligament disruption associated with odontoid fractures. Spine 1994;19:2307–2314.
41. Dickman CA, Greene KA, Sonntag VK. Injuries involving the transverse atlantal ligament: Classification and treatment guidelines based upon experience with 39 injuries. Neurosurgery 1996;38:44–50.
42. Heinrich SD, Gallagher D, Harris M, et al. Undiagnosed fractures in severely injured children and young adults. Identification with technetium imaging. J Bone Joint Surg Am 1994;76:561–572.
43. Bloom AI, Neeman Z, Slasky BS, et al. Fracture of the occipital condyles and associated craniocervical ligament injury: Incidence, CT imaging and implications. Clin Radiol 1997;52:198–202.
44. Kuzma BB, Goodman JM. Diagnosis of atlanto-occipital dislocation. Surg Neurol 1997;48:418–419.
45. Spence KF Jr, Decker S, Sell KW. Bursting atlantal fracture associated with rupture of the transverse ligament. J Bone Joint Surg Am 1970;52:543–549.
46. Fielding JW, Hawkins RJ, Hensinger RN, et al. Atlantoaxial rotary deformities. Orthop Clin North Am 1978;9:955–967.
47. Dvorak J, Hayek J, Zehnder R. CT-functional diagnostics of the rotatory instability of the upper cervical spine. Part 2. An evaluation on healthy adults and patients with suspected instability. Spine 1987;12:726–731.
48. Dvorak J, Panjabi M, Gerber M, et al. CT-functional diagnostics of the rotatory instability of upper cervical spine. 1. An experimental study on cadavers. Spine 1987;12:197–205.
49. Niibayashi H. Atlantoaxial rotatory dislocation. A case report. Spine 1998;23:1494–1496.
50. Pepin JW, Bourne RB, Hawkins RJ. Odontoid fractures, with special reference to the elderly patient. Clin Orthop 1985:178–183.
51. Levine AM, Edwards CC. Traumatic lesions of the occipitoatlantoaxial complex. Clin Orthop 1989:53–68.
52. Saternus KS. [Forms of fractures of the occipital condyles]. Z Rechtsmed 1987;99:95–108.
53. Anderson PA, Montesano PX. Morphology and treatment of occipital condyle fractures. Spine 1988;13:731–736.
54. Song WS, Chiang YH, Chen CY, et al. A simple method for diagnosing traumatic occlusion of the vertebral artery at the craniovertebral junction. Spine 1994;19:837–839.
55. Traynelis VC, Marano GD, Dunker RO, et al. Traumatic atlanto-occipital dislocation. Case report. J Neurosurg 1986;65:863–870.
56. Heller JG, Viroslav S, Hudson T. Jefferson fractures: The role of magnification artifact in assessing transverse ligament integrity. J Spinal Disord 1993;6:392–396.
57. Dickman CA, Mamourian A, Sonntag VK, et al. Magnetic resonance imaging of the transverse atlantal ligament for the evaluation of atlantoaxial instability. J Neurosurg 1991;75:221–227.
58. Levine AM, Edwards CC. Fractures of the atlas. J Bone Joint Surg Am 1991;73:680–691.
59. Mohit AA, Schuster JA, Mirza SK, et al. “Plough” fracture: Shear fracture of the anterior arch of the atlas. AJR Am J Roentgenol 2003;181:770.
60. Kesterson L, Benzel E, Orrison W, et al. Evaluation and treatment of atlas burst fractures (Jefferson fractures). J Neurosurg 1991;75:213–220.
61. Jefferson G. Remarks on fractures of the first cervical vertebra. Br Med J 1927;2:153–157.
62. Fielding JW, Hawkins RJ. Atlanto-axial rotatory fixation (fixed rotatory subluxation of the atlanto-axial joint). J Bone Joint Surg Am 1977;59:37–44.
63. Levine AM. Avulsion of the transverse ligament associated with a fracture of the atlas. Orthopedics 1983;6:1467–1471.
64. Weiner BK, Brower RS. Traumatic vertical atlantoaxial instability in a case of atlanto-occipital coalition. Spine 1997;22:1033–1035.
65. Hosono N, Yonenobu K, Kawagoe K, et al. Traumatic anterior atlanto-occipital dislocation. A case report with survival. Spine 1993;18:786–790.
66. Altongy JF, Fielding JW. Combined atlanto-axial and occipito-atlantal rotatory subluxation. A case report. J Bone Joint Surg Am 1990;72:923–926.
67. Ryan MD, Henderson JJ. The epidemiology of fractures and fracture-dislocations of the cervical spine. Injury 1992;23:38–40.
68. Anderson LD, D’Alonzo RT. Fractures of the odontoid process of the axis. J Bone Joint Surg Am 1974;56:1663–1674.
69. Hadley MN, Browner CM, Liu SS, et al. New subtype of acute odontoid fractures (type IIA). Neurosurgery 1988;22:67–71.
70. Chutkan NB, King AG, Harris MB. Odontoid fractures: Evaluation and management. J Am Acad Orthop Surg 1997;5:199–204.
71. Grauer JN, Shafi B, Hilibrand AS, et al. Proposal of a modified, treatment-oriented classification of odontoid fractures. Spine J 2005;5:123–129.
72. Bucholz RW. Unstable hangman’s fractures. Clin Orthop Relat Res 1981:119–124.
73. Effendi B, Roy D, Cornish B, et al. Fractures of the ring of the axis. A classification based on the analysis of 131 cases. J Bone Joint Surg Br 1981;63-B:319–327.
74. Starr JK, Eismont FJ. Atypical hangman’s fractures. Spine 1993;18:1954–1957.
75. Levine AM, Edwards CC. The management of traumatic spondylolisthesis of the axis. J Bone Joint Surg Am 1985;67:217–226.
76. Francis WR, Fielding JW, Hawkins RJ, et al. Traumatic spondylolisthesis of the axis. J Bone Joint Surg Br 1981;63-B:313–318.
77. Dickman CA, Hadley MN, Browner C, et al. Neurosurgical management of acute atlas-axis combination fractures. A review of 25 cases. J Neurosurg 1989;70:45–49.
78. Anderson PA, Budorick TE, Easton KB, et al. Failure of halo vest to prevent in vivo motion in patients with injured cervical spines. Spine 1991;16:S501–505.
79. Polin RS, Szabo T, Bogaev CA, et al. Nonoperative management of types II and III odontoid fractures: The Philadelphia collar versus the halo vest. Neurosurgery 1996;38:450–456; discussion 456–457.
80. Johnson RM, Hart DL, Simmons EF, et al. Cervical orthoses. A study comparing their effectiveness in restricting cervical motion in normal subjects. J Bone Joint Surg Am 1977;59:332–339.
81. Govender S, Grootboom M. Fractures of the dens—the results of non-rigid immobilization. Injury 1988;19:165–167.
82. Whitehill R, Richman JA, Glaser JA. Failure of immobilization of the cervical spine by the halo vest. A report of five cases. J Bone Joint Surg Am 1986;68:326–332.
83. Lind B, Nordwall A, Sihlbom H. Odontoid fractures treated with halo-vest. Spine 1987;12:173–177.
84. Schiff DC, Parke WW. The arterial supply of the odontoid process. J Bone Joint Surg Am 1973;55:1450–1456.
85. Guiot B, Fessler RG. Complex atlantoaxial fractures. J Neurosurg 1999;91:139–143.
86. Mirza SK, Moquin RR, Anderson PA, et al. Stabilizing properties of the halo apparatus. Spine 1997;22:727–733.
87. Bucci MN, Dauser RC, Maynard FA, et al. Management of post-traumatic cervical spine instability: Operative fusion versus halo vest immobilization. Analysis of 49 cases. J Trauma 1988;28:1001–1006.
88. Garfin SR, Botte MJ, Waters RL, et al. Complications in the use of the halo fixation device. J Bone Joint Surg Am 1986;68:320–325.
89. Glaser JA, Whitehill R, Stamp WG, et al. Complications associated with the halo-vest. A review of 245 cases. J Neurosurg 1986;65:762–769.
90. Schweigel JF. Halo-thoracic brace management of odontoid fractures. Spine 1979;4:192–194.
91. Bednar DA, Parikh J, Hummel J. Management of type II odontoid process fractures in geriatric patients; a prospective study of sequential cohorts with attention to survivorship. J Spinal Disord 1995;8:166–169.
92. Schlicke LH, Callahan RA. A rational approach to burst fractures of the atlas. Clin Orthop 1981:18–21.
93. Segal LS, Grimm JO, Stauffer ES. Non-union of fractures of the atlas. J Bone Joint Surg Am 1987;69:1423–1434.
94. Zimmerman E, Grant J, Vise WM, et al. Treatment of Jefferson fracture with a halo apparatus. Report of two cases. J Neurosurg 1976;44:372–375.
95. Han SY, Witten DM, Mussleman JP. Jefferson fracture of the atlas. Report of six cases. J Neurosurg 1976;44:368–371.
96. McGuire RA Jr, Harkey HL. Primary treatment of unstable Jefferson’s fractures. J Spinal Disord 1995;8:233–236.
97. Goel AL, Laheri V. Plate and screw fixation for atlanto-axial dislocation [technical report]. Acta Neurochir (Wien) 1994;129:47–53.
98. Fielding JW, Cochran GB, Lawsing JF 3rd, et al. Tears of the transverse ligament of the atlas. A clinical and biomechanical study. J Bone Joint Surg Am 1974;56:1683–1691.
99. Ryan MD, Taylor TK. Odontoid fractures. A rational approach to treatment. J Bone Joint Surg Br 1982;64:416–421.
100. Graziano G, Colon G, Hensinger R. Complete atlanto-axial dislocation associated with type II odontoid fracture: A report of two cases. J Spinal Disord 1994;7:518–521.
101. Apuzzo ML, Heiden JS, Weiss MH, et al. Acute fractures of the odontoid process. An analysis of 45 cases. J Neurosurg 1978;48:85–91.
102. Hadley MN, Dickman CA, Browner CM, et al. Acute axis fractures: A review of 229 cases. J Neurosurg 1989;71:642–647.
103. Dickson H, Engel S, Blum P, et al. Odontoid fractures, systemic disease and conservative care. Aust N Z J Surg 1984;54:243–247.
104. Dunn ME, Seljeskog EL. Experience in the management of odontoid process injuries: An analysis of 128 cases. Neurosurgery 1986;18:306–310.
105. Roy-Camille R, Saillant G, Judet T, et al. [Factors of severity in the fractures of the odontoid process (author’s transl)]. Rev Chir Orthop Reparatrice Appar Mot 1980;66:183–186.
106. Bohler J. Anterior stabilization for acute fractures and non-unions of the dens. J Bone Joint Surg Am 1982;64:18–27.
107. Nakanishi T, Sasaki T, Tokita N, et al. Internal fixation of the odontoid fracture. Orthop Trans 1982;6:176.
P.1495

108. Jeanneret B, Vernet O, Frei S, et al. Atlantoaxial mobility after screw fixation of the odontoid: A computed tomographic study. J Spinal Disord 1991;4:203–211.
109. Jeanneret B, Magerl F. Primary posterior fusion C1/2 in odontoid fractures: Indications, technique, and results of transarticular screw fixation. J Spinal Disord 1992;5:464–475.
110. Aebi M, Etter C, Coscia M. Fractures of the odontoid process. Treatment with anterior screw fixation. Spine 1989;14:1065–1070.
111. Apfelbaum RI, Lonser RR, Veres R, et al. Direct anterior screw fixation for recent and remote odontoid fractures. J Neurosurg 2000;93:227–236.
112. Montesano PX, Anderson PA, Schlehr F, et al. Odontoid fractures treated by anterior odontoid screw fixation. Spine 1991;16:S33–37.
113. Etter C, Coscia M, Jaberg H, et al. Direct anterior fixation of dens fractures with a cannulated screw system. Spine 1991;16:S25–32.
114. Przybylski GJ, Welch WC. Longitudinal atlantoaxial dislocation with type III odontoid fracture. Case report and review of the literature. J Neurosurg 1996;84:666–670.
115. Hadley MN, Sonntag VK, Grahm TW, et al. Axis fractures resulting from motor vehicle accidents. The need for occupant restraints. Spine 1986;11:861–864.
116. Cornish BL. Traumatic spondylolisthesis of the axis. J Bone Joint Surg Br 1968;50:31–43.
117. Norrell H, Wilson CB. Early anterior fusion for injuries of the cervical portion of the spine. JAMA 1970;214:525–530.
118. Tuite GF, Papadopoulos SM, Sonntag VK. Caspar plate fixation for the treatment of complex hangman’s fractures. Neurosurgery 1992;30:761–764; discussion 764–765.
119. Roda JM, Castro A, Blazquez MG. Hangman’s fracture with complete dislocation ofC-2 on C-3. Case report. J Neurosurg 1984;60:633–635.
120. Goto S, Tanno T, Moriya H. Cervical myelopathy caused by pseudoarthrosis between the atlas and axis associated with diffuse idiopathic skeletal hyperostosis. Spine 1995;20:2572–2575.
121. Sasso R, Doherty BJ, Crawford MJ, et al. Biomechanics of odontoid fracture fixation. Comparison of the one- and two-screw technique. Spine 1993;18:1950–1953.
122. Roy-Camille RS, Bouchet T. Technique du visage des pedicules de C2. In: Roy-Camille, ed. Cinquiemes Journees d’Orthopedie de la Pitie, Rachis Cervicale Superieur. Paris: Masson; 1986:41–43.
123. Solanki GA, Crockard HA. Peroperative determination of safe superior transarticular screw trajectory through the lateral mass. Spine 1999;24:1477–1482.
124. Reindl R, Sen M, Aebi M. Anterior instrumentation for traumatic C1–C2 instability. Spine 2003;28:E329–333.
125. Smith GW, Robinson RA. The treatment of certain cervical-spine disorders by anterior removal of the intervertebral disc and interbody fusion. J Bone Joint Surg Am 1958;40-A:607–624.
126. McAfee PC, Bohlman HH, Riley LH Jr, et al. The anterior retropharyngeal approach to the upper part of the cervical spine. J Bone Joint Surg Am 1987;69:1371–1383.
127. Vaccaro AR, Ring D, Lee RS, et al. Salvage anterior C1–C2 screw fixation and arthrodesis through the lateral approach in a patient with a symptomatic pseudoarthrosis. Am J Orthop 1997;26:349–353.
128. Dvorak J, Panjabi MM. Functional anatomy of the alar ligaments. Spine 1987;12:183–189.
129. Zavanone M, Guerra P, Rampini P, et al. Traumatic fractures of the craniovertebral junction. Management of 23 cases. J Neurosurg Sci 1991;35:17–22.
130. Fairholm D, Lee ST, Lui TN. Fractured odontoid: The management of delayed neurological symptoms. Neurosurgery 1996;38:38–43.
131. Magerl FSCS. Stable posterior fusion of the atlas and axis by transarticular screw fixation. In: Kehr PW, Weidner A, eds. Cervical spine. Berlin: Springer-Verlag; 1986:322–327.
132. Harms J, Melcher RP. Posterior C1–C2 fusion with polyaxial screw and rod fixation. Spine 2001;26:2467–2471.
133. Nadim Y, Lu J, Sabry FF, et al. Occipital screws in occipitocervical fusion and their relation to the venous sinuses: An anatomic and radiographic study. Orthopedics 2000;23:717–719.
134. Brooks AL, Jenkins EB. Atlanto-axial arthrodesis by the wedge compression method. J Bone Joint Surg Am 1978;60:279–284.
135. Griswold DM, Albright JA, Schiffman E, et al. Atlanto-axial fusion for instability. J Bone Joint Surg Am 1978;60:285–292.
136. Gallie WE. Fractures and dislocations of the cervical spine. Am J Surg 1939;46:494–499.
137. Grob D, Magerl F. [Surgical stabilization of C1 and C2 fractures]. Orthopade 1987;16:46–54.
138. Madawi AA, Casey AT, Solanki GA, et al. Radiological and anatomical evaluation of the atlantoaxial transarticular screw fixation technique. J Neurosurg 1997;86:961–968.
139. Paramore CG, Dickman CA, Sonntag VK. The anatomical suitability of the C1–C2 complex for transarticular screw fixation. J Neurosurg 1996;85:221–224.
140. Tokuda K, Miyasaka K, Abe H, et al. Anomalous atlantoaxial portions of vertebral and posterior inferior cerebellar arteries. Neuroradiology 1985;27:410–413.
141. Wright NM. Posterior C2 fixation using bilateral, crossing c2 laminar screws: Case series and technical note. J Spinal Disord Tech 2004;17:158–162.
142. Ebraheim NA, Misson JR, Xu R, et al. The optimal transarticular C1–C2 screw length and the location of the hypoglossal nerve. Surg Neurol 2000;53:208–210.
143. Nadim Y, Sabry F, Xu R, et al. Computed tomography in the determination of transarticular C1–C2 screw length. Orthopedics 2000;23:373–375.
144. Gebhard JS, Schimmer RC, Jeanneret B. Safety and accuracy of transarticular screw fixation C1–C2 using an aiming device. An anatomic study. Spine 1998;23:2185–2189.
145. Hanson PB, Montesano PX, Sharkey NA, et al. Anatomic and biomechanical assessment of transarticular screw fixation for atlantoaxial instability. Spine 1991;16:1141–1145.
146. Melcher RP, Puttlitz CM, Kleinstueck FS, et al. Biomechanical testing of posterior atlantoaxial fixation techniques. Spine 2002;27:2435–2440.
147. Goel A. C1–C2 pedicle screw fixation with rigid cantilever beam construct: Case report and technical note. Neurosurgery 2002;51:853–854; author reply 854.
148. Young JP, Young PH, Anderson PA, et al. The ponticulus posticus: Implications for C1 lateral mass screws. 32nd Annual Meeting of the Cervical Spine Research Society. Boston, MA; 2004.
149. Resnick DK, Lapsiwala S, Trost GR. Anatomic suitability of the C1–C2 complex for pedicle screw fixation. Spine 2002;27:1494–1498.
150. Sagi HC, Beutler W, Carroll E, et al. Airway complications associated with surgery on the anterior cervical spine. Spine 2002;27:949–953.
151. Bell C. Surgical observations. Middlesex Hosp J 1817;4:469–470.
152. Noble ER, Smoker WR. The forgotten condyle: The appearance, morphology, and classification of occipital condyle fractures. AJNR Am J Neuroradiol 1996;17:507–513.
153. Urculo E, Arrazola M, Arrazola M Jr, et al. Delayed glossopharyngeal and vagus nerve paralysis following occipital condyle fracture. Case report. J Neurosurg 1996;84:522–525.
154. Grabb BC, Frye TA, Hedlund GL, et al. MRI diagnosis of suspected atlanto-occipital dissociation in childhood. Pediatr Radiol 1999;29:275–281.
155. Stroobants J, Fidlers L, Storms JL, et al. High cervical pain and impairment of skull mobility as the only symptoms of an occipital condyle fracture. Case report. J Neurosurg 1994;81:137–138.
156. Bundschuh CV, Alley JB, Ross M, et al. Magnetic resonance imaging of suspected atlanto-occipital dislocation. Two case reports. Spine 1992;17:245–248.
157. Dickman CA, Papadopoulos SM, Sonntag VK, et al. Traumatic occipitoatlantal dislocations. J Spinal Disord 1993;6:300–313.
158. Ferrera PC, Bartfield JM. Traumatic atlanto-occipital dislocation: A potentially surviv-able injury. Am J Emerg Med 1996;14:291–296.
159. Montane I, Eismont FJ, Green BA. Traumatic occipitoatlantal dislocation. Spine 1991;16:112–116.
160. Zielinski CJ, Gunther SF, Deeb Z. Cranial-nerve palsies complicating Jefferson fracture. A case report. J Bone Joint Surg Am 1982;64:1382–1384.
161. Krantz P. Isolated disruption of the transverse ligament of the atlas: An injury easily overlooked at post-mortem examination. Injury 1980;12:168–170.
162. Wigren A, Sweden U, Amici F Jr. Traumatic atlanto-axial dislocation without neurological disorder. A case report. J Bone Joint Surg Am 1973;55:642–644.
163. Geisler FH, Cheng C, Poka A, et al. Anterior screw fixation of posteriorly displaced type II odontoid fractures. Neurosurgery 1989;25:30–37; discussion 37–38.
164. Crockard HA, Heilman AE, Stevens JM. Progressive myelopathy secondary to odontoid fractures: Clinical, radiological, and surgical features. J Neurosurg 1993;78:579–586.
165. Schatzker J, Rorabeck CH, Waddell JP. Fractures of the dens (odontoid process). An analysis of thirty-seven cases. J Bone Joint Surg Br 1971;53:392–405.
166. Hensinger RN, Fielding JW, Hawkins RJ. Congenital anomalies of the odontoid process. Orthop Clin North Am 1978;9:901–912.
167. Greene KA, Dickman CA, Marciano FF, et al. Acute axis fractures. Analysis of management and outcome in 340 consecutive cases. Spine 1997;22:1843–1852.
168. Alander DH, Andreychik DA, Stauffer ES. Early outcome in cervical spinal cord injured patients older than 50 years of age. Spine 1994;19:2299–2301.
169. Southwick WO. Management of fractures of the dens (odontoid process). J Bone Joint Surg Am 1980;62:482–486.
170. Borne GM, Bedou GL, Pinaudeau M, et al. Odontoid process fracture osteosynthesis with a direct screw fixation technique in nine consecutive cases. J Neurosurg 1988;68:223–226.
171. Schweigel JF. Management of the fractured odontoid with halo-thoracic bracing. Spine 1987;12:838–839.
172. Haid RW Jr, Subach BR, McLaughlin MR, et al. C1–C2 transarticular screw fixation for atlantoaxial instability: A 6-year experience. Neurosurgery 2001;49:65–68; discussion 69–70.
173. Spivak JM, Weiss MA, Cotler JM, et al. Cervical spine injuries in patients 65 and older. Spine 1994;19:2302–2306.
174. Lieberman IH, Webb JK. Cervical spine injuries in the elderly. J Bone Joint Surg Br 1994;76:877–881.
175. Friedman D, Flanders A, Thomas C, et al. Vertebral artery injury after acute cervical spine trauma: Rate of occurrence as detected by MR angiography and assessment of clinical consequences. AJR Am J Roentgenol 1995;164:443–447; discussion 448–449.
176. Renwick IG. The type III dens fracture and its associated soft-tissue injuries: A different form of hangman’s fracture. Br J Radiol 1990;63:495–496.
177. Sherk HH, Snyder B. Posterior fusions of the upper cervical spine: Indications, techniques, and prognosis. Orthop Clin North Am 1978;9:1091–1099.
178. Lipscomb PR. Cervico-occipital fusion for congenital and post-traumatic anomalies of the atlas and axis. J Bone Joint Surg Am 1957;39-A:1289–1301.
179. Grantham SA, Dick HM, Thompson RC Jr, et al. Occipitocervical arthrodesis. Indications, technic and results. Clin Orthop 1969;65:118–129.
180. Abumi K, Takada T, Shono Y, et al. Posterior occipitocervical reconstruction using cervical pedicle screws and plate-rod systems. Spine 1999;24:1425–1434.
181. Wertheim SB, Bohlman HH. Occipitocervical fusion. Indications, technique, and long-term results in thirteen patients. J Bone Joint Surg Am 1987;69:833–836.
182. Smith DC. Atlanto-occipital dislocation. J Emerg Med 1992;10:699–703.
183. Elia M, Mazzara JT, Fielding JW. Onlay technique for occipitocervical fusion. Clin Orthop 1992:170–174.
184. Newman P, Sweetnam R. Occipito-cervical fusion. An operative technique and its indications. J Bone Joint Surg Br 1969;51:423–431.
185. Haher TR, Yeung AW, Caruso SA, et al. Occipital screw pullout strength. A biomechanical investigation of occipital morphology. Spine 1999;24:5–9.
186. Hamblen DL. Occipito-cervical fusion. Indications, technique and results. J Bone Joint Surg Br 1967;49:33–45.
P.1496

187. Ransford AO, Crockard HA, Pozo JL, et al. Craniocervical instability treated by contoured loop fixation. J Bone Joint Surg Br 1986;68:173–177.
188. Itoh T, Tsuji H, Katoh Y, et al. Occipito-cervical fusion reinforced by Luque’s segmental spinal instrumentation for rheumatoid diseases. Spine 1988;13:1234–1238.
189. Sasso RC, Jeanneret B, Fischer K, et al. Occipitocervical fusion with posterior plate and screw instrumentation. A long-term follow-up study. Spine 1994;19:2364–2368.
190. Smith MD, Anderson P, Grady MS. Occipitocervical arthrodesis using contoured plate fixation. An early report on a versatile fixation technique. Spine 1993;18:1984–1990.
191. Hajek PD, Lipka J, Hartline P, et al. Biomechanical study of C1–C2 posterior arthrodesis techniques. Spine 1993;18:173–177.
192. Grob D, Crisco JJ 3rd, Panjabi MM, et al. Biomechanical evaluation of four different posterior atlantoaxial fixation techniques. Spine 1992;17:480–490.
193. Fried LC. Atlanto-axial fracture-dislocations. Failure of posterior C.1 to C.2 fusion. J Bone Joint Surg Br 1973;55:490–496.
194. McGraw RW, Rusch RM. Atlanto-axial arthrodesis. J Bone Joint Surg Br 1973;55:482–489.
195. Fielding JW, Hawkins RJ, Ratzan SA. Spine fusion for atlanto-axial instability. J Bone Joint Surg Am 1976;58:400–407.
196. Bednar DA. Nonunion of cervical burst fracture: A unique observation. J Spinal Disord 1990;3:384–386.
197. Wright NM, Lauryssen C. Vertebral artery injury in C1–C2 transarticular screw fixation: Results of a survey of the AANS/CNS section on disorders of the spine and peripheral nerves. American Association of Neurological Surgeons/Congress of Neurological Surgeons. J Neurosurg 1998;88:634–640.
198. Grob D, Jeanneret B, Aebi M, et al. Atlanto-axial fusion with transarticular screw fixation. J Bone Joint Surg Br 1991;73:972–976.
199. Vaccaro AR, Lehman AP, Ahlgren BD, et al. Anterior C1–C2 screw fixation andbony fusion through an anterior retropharyngeal approach. Orthopedics 1999;22:1165–1170.
200. Khodadadyan-Klostermann C, Kandziora F, Schnake KJ, et al. [Transoral atlanto-axial plate fixation in the treatment of a malunited dens fracture and secondary atlanto-axial instability]. Chirurg 2001;72:1298–1302.
201. Panjabi M, Dvorak J, Duranceau J, et al. Three-dimensional movements of the upper cervical spine. Spine 1988;13:726–730.
202. Bazaz R, Lee MJ, Yoo JU. Incidence of dysphagia after anterior cervical spine surgery: A prospective study. Spine 2002;27:2453–2458.
203. Henry AD, Bohly J, Grosse A. Fixation of odontoid fractures by an anterior screw. J Bone Joint Surg Br 1999;81:472–477.
204. Borm W, Kast E, Richter HP, et al. Anterior screw fixation in type II odontoid fractures: Is there a difference in outcome between age groups? Neurosurgery 2003;52:1089–1092; discussion 1084–1092.
205. Andersson S, Rodrigues M, Olerud C. Odontoid fractures: High complication rate associated with anterior screw fixation in the elderly. Eur Spine J 2000;9:56–59; discussion 60.
206. Graziano G, Jaggers C, Lee M, et al. A comparative study of fixation techniques for type II fractures of the odontoid process. Spine 1993;18:2383–2387.
207. Daentzer D, Deinsberger W, Boker DK. Vertebral artery complications in anterior approaches to the cervical spine: Report of two cases and review of literature. Surg Neurol 2003;59:300–309; discussion 309.