Rockwood & Wilkins’ Fractures in Children
6th Edition

Chapter 18
Cervical Spine Injuries in Children
William C. Warner Jr
Daniel J. Hedequist
Cervical spine fractures in children are rare, accounting for only 1% of pediatric fractures and 2% of all spinal injuries (1,2,3,4,5,6,7,8,9,10,11). The incidence is estimated to be 7.41 in 100,000 per year (12); however, that may be misleading because some injuries are not detected or are detected only at autopsy. Aufdermaur (13) examined the autopsied spines of 12 juveniles who had spinal injuries. All 12 had cartilage endplates that were separated from the vertebral bodies in the zone of columnar and calcified cartilage, similar to a Salter I fracture, although clinically and by x-ray a fracture was suggested in only 1 patient. Only x-rays at autopsy showed the disruption, represented by a small gap or apparent widening of the intervertebral space (13).
Cervical spine injuries in children younger than 8 years of age occur in the lower cervical spine, while older children and

adolescents tend to have fractures involving either the upper or lower cervical spine (14). The upper cervical spine in children is more prone to injury because of the anatomic and biomechanical properties of the immature spine. The immature spine is hypermobile because of ligamentous laxity, and the facet joints are oriented in a more horizontal position; both of these properties predispose children to more forward translation. Younger children also have a relatively large head compared to the body, which changes the fulcrum of motion of the upper cervical spine. All of the factors predispose younger children to injuries of the upper cervical spine; with age, the anatomic changes lead to an increased prevalence of lower cervical spine injuries.
The mechanism of injury in pediatric patients also is age-related. Infants with cervical spine injuries should be evaluated for abuse (15). In children up to 9 years of age, the most common mechanism of injury is related to motor vehicle accidents; the second most common mechanism in this age group is falls (16,17). In older children and adolescents, the most common mechanism of cervical spine injuries is sporting activities, followed by motor vehicle accidents (16,18).
Cervical spine injuries associated with neurologic deficits are infrequent in children, and when incomplete there tends to be a better prognosis for recovery in children than in adults (19,20). Complete neurologic deficits, regardless of patient age, tend to have a poor prognosis for any recovery and may be indicative of the severity and magnitude of injury (21,22,23,24). Death from cervical spine injuries tends to be related to the level of injury and the associated injuries. Higher cervical spine injuries (i.e., atlanto-occipital dislocation) in younger children are associated with the highest mortality rate (25,26). Children with significant cervical spine injuries also may have associated severe head injuries, leading to an increase in mortality. In a study of 61 pediatric deaths related to spinal cord injuries, 89% of fatalities occurred at the scene, and most were related to high cervical cord injuries in patients who had sustained multiple injuries (27).
Understanding the normal growth and development of the cervical spine is essential when treating a child with a suspected cervical spine injury. This will allow the physician to differentiate normal physes or synchondroses from pathologic fractures or ligamentous disruptions and will alert the physician to any possible congenital anomalies that may be mistaken for a fracture.
Upper Cervical Spine
At birth the atlas is composed of three ossification centers, one for the body and one for each of the neural arches (Fig. 18-1). The ossification center for the anterior arch is present in approximately 20% of individuals at birth, appearing in the remainder during the first year of life. Occasionally, the anterior arch is bifid, and the body may be formed from two centers or may fail to completely appear. The posterior arches usually fuse by the age of 3 years; however, occasionally the posterior synchondrosis between the two fails to fuse, resulting in a bifid arch. The neurocentral synchondroses that link the neural arches to the body are best seen on an open-mouth odontoid view. These synchondroses close by 7 years of age and should not be mistaken for fractures (28). The canal of the atlas is large to allow for the amount of rotation that occurs at this joint as well as some forward translation (29). The vertebral arteries are about 2 cm from the midline and run in a groove on the superior surface of the atlas. This must be remembered during lateral dissection at the occipital cervical junction. Because the ring of C1 reaches about normal adult size by 4 years of age, arthrodesis after this time should not cause spinal canal stenosis.
FIGURE 18-1 Diagram of C1 (atlas). The body (A) is not ossified at birth, and its ossification center appears during the first year of life. The body may fail to develop, and forward extension of neural arches (C) may take its place. Neural arches appear bilaterally about the 7th week (D), and the most anterior portion of the superior articulating surface usually is formed by the body. The synchondrosis of the spinous processes unites by the third year. Union rarely is preceded by the appearance of the secondary center within the synchondrosis. Neurocentral synchondrosis (F) fuses about the seventh year. The ligament surrounding the superior vertebral notch (K) may ossify, especially in later life. (Reprinted from Bailey DK. Normal cervical spine in infants and children. Radiology 1952;59:713–714; with permission.)
The axis develops from at least four separate ossification

centers: one for the dens, one for the body, and two for the neural arches (Fig. 18-2). Between the odontoid and the body of the axis is a synchondrosis or vestigial disk space that often is mistaken for a fracture line. This synchondrosis runs well below the level of the articular processes of the axis and usually fuses at 6 to 7 years of age, although it may persist as a sclerotic line until 11 years of age (29). The most common odontoid fracture pattern in adults and adolescents is transverse and at the level of the articular processes. The normal synchondrosis should not be confused with this fracture; the synchondrosis is more cup-shaped and below the level of the articular processes. After 7 years of age, the synchondrosis should not be present on an open-mouth odontoid view; a fracture should be considered if a lucent line is present after this age. The neural arches of C2 fuse at 3 to 6 years of age; these are seen as vertical lucent lines on the open-mouth odontoid view. Occasionally, the tip of the odontoid is V-shaped (dens bicornum), or a small separate summit ossification center may be present at the tip of the odontoid (ossiculum terminale). An os odontoideum is believed to result from a history of unrecognized trauma. The differentiation between an os odontoideum and the synchondrosis of the body is relatively easy because of their relationships to the level of the C1-C2 facet (Fig. 18-3).
FIGURE 18-2 Diagram of C2 (axis). The body (A) in which one center (occasionally two) appears by the fifth fetal month. Neural arches (C) appear bilaterally by the seventh fetal month. Neural arches fuse (D) posteriorly by the second or third year. Bifid tip (E) of spinous process (occasionally a secondary center is present in each tip). Neurocentral synchondrosis (F) fuses at 3 to 6 years. The inferior epiphyseal ring (G) appears at puberty and fuses at about 25 years of age. The summit ossification center (H) for the odontoid appears at 3 to 6 years and fuses with the odontoid by 12 years. Odontoid (dens) (I). Two separate centers appear by the fifth fetal month and fuse with each other by the seventh fetal month. The synchondrosis between the odontoid and neural arch (I) fuses at 3 to 6 years. Synchondrosis between the odontoid and body (L) fuses at 3 to 6 years. Posterior surface of the body and odontoid (M). (Reprinted from Bailey DK. Normal cervical spine in infants and children. Radiology 1952;59:713–714; with permission.)
The arterial supply to the odontoid is derived from the vertebral and carotid arteries. The anterior and posterior ascending arteries arise from the vertebral artery at the level of C3 and ascend anterior and posterior to the odontoid, meeting superiorly to form an apical arcade. These arteries supply small penetrating branches to the body of the axis and the odontoid process. The internal carotid artery gives off cleft perforators that supply the superior portion of the odontoid. This arrangement of arteries and vessels is necessary for embryologic development and anatomic function of the odontoid. The synchondrosis prevents direct vascularization of the odontoid from C2, and vascularization from the blood supply of C1 is not possible because the synovial joint cavity surrounds the odontoid. The formation of an os odontoideum after cervical trauma may be related to this peculiar blood supply (Fig. 18-4).
Lower Cervical Spine
The third through seventh cervical vertebrae share a similar ossification pattern: a single ossification center for the vertebral body and an ossification center for each neural arch (Fig. 18-5). The neural arch fuses posteriorly between the second and third years, and the neurocentral synchondroses between the neural arches and the vertebral body fuse by 3 to 6 years of

age. These vertebrae normally are wedge-shaped until 7 to 8 years of age (13,30,31). The vertebral bodies, neural arches, and pedicles enlarge by periosteal appositional growth, similar to that seen in long bones. By 8 to 10 years of age, a child’s spine usually reaches near adult size and characteristics. There are five secondary ossification centers that can remain open until 25 years of age (30). These include one each for the spinous processes, transverse processes, and the ring apophyses about the vertebral endplates. These should not be confused with fractures.
FIGURE 18-3 CT scan showing presence of an os odontoideum. Note the position of the os well above the C1-C2 facets. The scan also shows the vestigial scar of the synchondrosis between the dens and the body below the C1-C2 facet.
FIGURE 18-4 Blood supply to odontoid: posterior and anterior ascending arteries and apical arcade. (Reprinted from Schiff DC, Parke WW. The arterial supply of the odontoid process. J Bone Joint Surg [Am] 1973;55:1450–1464; with permission.)
The superior and inferior endplates are firmly bound to the adjacent disk. The junction between the vertebral body and the endplate is similar to a physis of a long bone. The vertebral body is analogous to the metaphysis and the endplate to the physis, where longitudinal growth occurs. The junction between the vertebral body and the endplate has been shown to be weaker than the adjacent vertebral body or disk, which can result in a fracture at the endplate in the area of columnar and calcified cartilage of the growth zone, similar to a Salter-Harris I fracture of a long bone (13). The inferior endplate may be more susceptible to this injury than the superior endplate because of the mechanical protection afforded by the developing uncinate processes (32).
FIGURE 18-5 Diagram of typical cervical vertebrae, C3 to C7. The body (A) appears by the fifth fetal month. The anterior (costal) portion of the transverse process (B) may develop from a separate center that appears by the sixth fetal month and joins the arch by the sixth year. Neural arches (C) appear by the seventh to ninth fetal week. The synchondrosis between spinous processes (D) usually unites by the second or third year. Secondary centers for bifid spine (E) appear at puberty and unite with spinous process at 25 years. Neurocentral synchondrosis (F) fuses at 3 to 6 years. Superior and inferior epiphyseal rings (G) appear at puberty and unite with the body at about 25 years. The seventh cervical vertebra differs slightly because of a long, powerful, nonbifid spinous process. (Reprinted from Bailey DK. Normal cervical spine in infants and children. Radiology 1952;59:713–714; with permission.)
The facet joints of the cervical spine change in orientation with age. The angle of the C1-C2 facet is 55 degrees in newborns and increases to 70 degrees at maturity. In the lower cervical spine, the angle of the facet joints is 30 degrees at birth and 60 to 70 degrees at maturity. This may explain why the pediatric cervical spine may be more susceptible to injury from the increased motion or translation allowed by the facet joint orientation.

Increased ligamentous laxity in young children allows a greater degree of spinal mobility than in adults. Flexion and extension of the spine at C2-C3 are 50% greater in children between the ages of 3 and 8 years than in adults. The level of the greatest mobility in the cervical spine descends with increasing age. Between 3 and 8 years of age the most mobile segment is C3-C4; from 9 to 11 years, C4-C5 is the most mobile segment, and from 12 to 15 years, C5-C6 is the most mobile segment (33,34). This explains the tendency for craniocervical injuries in the young children.
Several anomalies of the cervical spine may influence treatment recommendations. The atlas can fail to segment from the skull, a condition called occipitalization of the atlas, and can lead to narrowing of the foramen magnum, neurologic symptoms, and increased stresses to the atlantoaxial articulation, which often causes instability. Failure of fusion of the posterior arch of C1 is not uncommon and should be sought before any procedure that involves C1. Wedge-shaped vertebrae, bifid vertebrae, or a combination of these also can occur. Klippel-Feil syndrome consists of the classic triad of a short neck, low posterior hairline, and severe restriction of motion of the neck from fusion of the cervical vertebrae (35,36). Congenital fusion of the cervical spine may predispose a child to injury from trauma by concentrating stresses in the remaining mobile segments.
Hensinger et al (37) reported congenital anomalies of the odontoid, including aplasia (complete absence), hypoplasia (partial absence in which there is a stubby piece at the base of the odontoid located above the C1 articulation), and os odontoideum. Os odontoideum consists of a separate ossicle of the odontoid with no connection to the body of C2. The cause may be traumatic. These anomalies also may predispose a child to injury or instability.
Most cervical spine injuries in young children are the result of motor vehicle accidents, sporting injuries, or pedestrian injuries (38). Infants are at risk for cervical spine injuries during the obstetric period, as well as during early development because of their lack of head control; however, most cervical spine injuries in infants are spinal cord injury without radiologic abnormality (SCIWORA) and are related to child abuse (13). Younger children may sustain injuries to their neck from seemingly low-energy falls of less than 5 feet; however, most of their cervical spine injuries are sustained as a result of motor vehicle accidents (39,40). As children become adolescents the prevalence of sporting injuries increases, as does the prevalence of athletic-related SCIWORA (16,38). Regardless of the cause, an adequate history may be difficult to obtain at the initial evaluation, and repeat evaluations may be needed.
The most common presenting symptom in patients with cervical spine injuries is pain localized to the cervical region. Other complaints, such as headache, inability to move the neck, subjective feelings of instability, and neurologic symptoms, all warrant complete evaluation. Infants may present with unexplained respiratory distress, motor weakness, or hypotonia, which warrant further evaluation. Patients with head and neck trauma, distracting injuries, or altered levels of consciousness are at high risk for a cervical spine injury and need to be thoroughly evaluated before obtaining cervical spine clearance (16). The presence of an occult cervical spine injury in an uncooperative or obtunded patient needs to be considered because of the frequency of SCIWORA in the pediatric population (9,10).
The evaluation of any patient with a suspected cervical spine injury should begin with inspection. Head and neck trauma is associated with a high incidence of cervical spine injuries (13,33). Soft tissue abrasions or shoulder-harness marks on the neck from a seatbelt are clues to an underlying cervical spine injury (Fig. 18-6) (15,41,42). Unconscious patients should be treated as if they have a cervical spine injury until further evaluation proves otherwise. The next step in the evaluation is palpation of the cervical spine for tenderness, muscle spasm, and overall alignment. The most prominent levels should be the spinous processes at C2, C3, and C7. Anterior palpation should focus on the presence of tenderness or swelling. The entire spine should be palpated and thoroughly examined because 20% of patients with cervical spine injuries have other spinal fractures.
A thorough neurologic examination should be done, which can be difficult in pediatric patients. Strength, sensation, reflexes, and proprioception should be documented. In patients who are uncooperative because of age or altered mental status, repeat examinations are important; however, the initial neurovascular

examination should be documented even it if entails only gross movements of the extremities. The evaluation of rectal sphincter tone, bulbocavernosus reflex, and perianal sensation are important, especially in obtunded patients and patients with partial or complete neurologic injuries, regardless of age. Patients who are cooperative and awake can be asked to perform supervised flexion, extension, lateral rotation, and lateral tilt. Uncooperative or obtunded patients should not have any manipulation of the neck.
FIGURE 18-6 Clinical photograph of a patient with a cervical spine injury resulting from impact with the shoulder harness of a seat belt. Note location of skin contusions from the seat belt.
Plain X-Rays
Plain x-rays remain the standard for evaluating the cervical spine in children. There currently is no consensus regarding whether or not all pediatric trauma patients require cervical spine films. The presence of tenderness and a distracting injury are the most common clinical presentations of a cervical spine injury (43). While some studies have shown that plain x-rays are of low yield in patients without evidence of specific physical findings, the burden remains on the treating physician to clear the cervical spine (44,45,46,47). Clearly, patients with tenderness, distracting injuries, neurologic deficits, head and neck trauma, and altered levels of consciousness need to have a complete set of cervical spine x-rays. Initial x-rays should include an anteroposterior view, open-mouth odontoid view, and lateral view of the cervical spine. Patients who are deemed unstable in the emergency room and are not able to tolerate multiple x-rays should have a cross-table lateral view of the cervical spine until further x-rays can be taken. The false-negative rates for a single cross-table x-ray have been reported to be 23% to 26%, indicating that complete x-rays are necessary when the patient is stable (48,49).
Flexion and extension x-rays may further aid the evaluation of the cervical spine, but these views are unlikely to be abnormal when standard views show no abnormalities. These views are helpful, however, in ruling out acute ligamentous injury (50). We recommend flexion and extension views in an alert patient with midline tenderness who has normal plain films of the cervical spine. These views must be taken only with a cooperative and alert child; they should not be used in obtunded or uncooperative patients, nor should they be done by manually placing the child in a position of flexion and extension.
Evaluation of cervical spine x-rays should proceed with a knowledge of the anatomic ossification centers and variations that occur in children. Each vertebral level should be systematically evaluated, as should the overall alignment of the cervical spine with respect to the anterior and posterior aspects of the vertebral bodies, the spinolaminar line, and the interspinous distances. The absence of cervical lordosis, an increase in the prevertebral soft tissue space, and subluxation of C2 on C3 are all anatomic variations that may be normal in children (28). Ossification centers also may be confused with fractures, most commonly in evaluation of the dens. The presence of a synchondrosis at the base of the odontoid can be distinguished from a fracture based on the age of the patient and the location of synchondrosis well below the facet joints. Knowledge of these normal variants is useful in evaluating plain x-rays of the cervical spine in children (Table 18-1).
X-Ray Evaluation of Specific Areas of the Spine
Atlanto-Occipital Junction
The atlanto-occipital interval remains the most difficult to assess for abnormalities, partly because of the difficulty in obtaining quality x-rays and partly because of the lack of discrete and reproducible landmarks. The distance between the occipital condyles and the facet joints of the atlas should be less than 5 mm; any distance of more than this suggests an atlanto-occipital disruption (34,51). The foramen magnum and its relationship to the atlas also are useful in detecting injuries of the atlanto-occipital region. The anterior cortical margin of the foramen magnum is termed the basion, while the posterior cortical margin of the foramen magnum is termed the opisthion. The distance between the basion and the tip of the dens should be less

than 12 mm as measured on a lateral x-ray (52). The Powers ratio (Fig. 18-7) is used to assess the position of the skull base relative to the atlas and is another way of evaluating the atlanto-occipital region. To determine this ratio, a line is drawn from the basion to the anterior cortex of the posterior arch of C1, and this distance is divided by the distance of a line drawn from the opisthion to the posterior cortex of the anterior arch of C1. The value should be between 0.7 and 1; a higher value indicates anterior subluxation of the atlanto-occipital joint and a lower value indicates a posterior subluxation. The problem lies in the fact that the basion is not always visible on plain x-rays. The Wackenheim line, which is drawn along the posterior aspect of the clivus, probably is the most easily identified line to determine disruption of the atlanto-occipital joint. If the line does not intersect the tip of the odontoid tangentially and if this line is displaced anteriorly or posteriorly, disruption or increased laxity about the atlanto-occipital joint should be suspected.
TABLE 18-1 Normal Ossification Centers and Anomalies Frequently Confused with Injury
Avulsion fracture
   Apical ossification center of the odontoid
   Secondary ossification centers at the tips of the transverse and spinous processes
   Persistence of the synchondrosis at the base of the odontoid
   Apparent anterior wedging of a young child’s vertebral body
   Normal posterior angulation of the odontoid seen in 4% of normal children
   Pseudosubluxation of C2-C3
   Incomplete ossification, especially of the odontoid process, with apparent superior subluxation of the anterior arch of C1
   Absence of the ossification center of the anterior arch of C1 in the first year of life may suggest posterior displacement of C1 on the odontoid
   Increase in the atlanto–dens interval of up to 4.5 mm
   Physiologic variations in the width of the prevertebral soft tissue due to crying misinterpreted as swelling due to edema or hemorrhage
   Overlying structures such as ears, braided hair, teeth, or hyoid bone. Plastic rivets used in modern emergency cervical immobilization collars can simulate fracture line.
   Horizontally placed facets in the younger child, creating the illusion of a pillar fracture
   Congenital anomalies such as os odontoideum, spina bifida, and congenital fusion or hemivertebrae
FIGURE 18-7 The Powers ratio is determined by drawing a line from the basion (B) to the posterior arch of the atlas (C) and a second line from the opisthion (O) to the anterior arch of the atlas (A). The length of the line BC is divided by the length of the line OA, producing the Powers ratio. (Reprinted from Lebwohl NH, Eismont FJ. Cervical spine injuries in children. In: Weinstein SL, ed. The pediatric spine: principles and practice. New York: Raven, 1994; with permission.)
FIGURE 18-8 The atlantodens interval (ADI) and the space available for cord (SAC) are used in determining atlantoaxial instability. The Wackenheim clivus-canal line is used to determine atlantooccipital injury, while the McRae and McGregor lines are used in the measurement of basilar impression. (Reprinted from Copley LA, Dormans JP. Cervical spine disorders in infants and children. J Am Acad Orthop Surg 1998;6:204–214; with permission.)
Atlantoaxial Joint
The atlanto–dens interval (ADI) and the space available for the spinal canal (SAC) are two useful measurements for evaluation of the atlantoaxial joint (Fig. 18-8). The ADI in a child is considered normal up to 4.5 mm, partly because the unossified cartilage of the odontoid, which is not seen on plain films, gives an apparent increase in the interval. At the level of the atlan-toaxial joint, the space taken up is broken into Steel’s rule of thirds: one third is taken up by the odontoid, one third by

the spinal cord, and one third is space available for the cord. These intervals also are easily measured on flexion and extension views and are helpful in determining instability. In children, extension views give the appearance of subluxation of the anterior portion of the atlas over the unossified dens, but this represents a pseudosubluxation and not instability (29,53).
FIGURE 18-9 The spinolaminar line (Swischuk line) is used to determine the presence of pseudosubluxation of C2 on C3. (Reprinted from Copley LA, Dormans JP. Cervical spine disorders in infants and children. J Am Acad Orthop Surg 1998;6:201–214; with permission.)
Upper Cervical Spine
Anterior displacement of one vertebral body on another may or may not indicate a true bony or ligamentous injury. Displacement of less than 3 mm at one level is a common anatomic variant in children at the levels of C2-C3 and C3-C4. This displacement is seen on flexion x-rays and reduces in extension. The posterior line of Swischuk (54) has been described to differentiate pathologic subluxation from normal anatomic variation; this line is drawn from the anterior cortex of the spinous process of C1 to the spinous process of C3 (Fig. 18-9). The anterior cortex of the spinous process of C2 should lie within 3 mm of this line; if the distance is more than this, a true subluxation should be suspected (Fig. 18-10). Widening of the spinous processes between C1 and C2 of more than 10 mm also is indicative of a ligamentous injury and should be evaluated by further imaging studies (55).
Lower Cervical Spine
Lateral x-rays of the cervical spine should be evaluated for overall alignment as well as at each level. The overall alignment can be evaluated by the continuous lines formed by the line adjoining the spinous processes, the spinolaminar line, and the lines adjoining the posterior and anterior vertebral bodies (Fig. 18-11). These lines should all be smooth and continuous with

no evidence of vertebral translation at any level. Loss of normal cervical lordosis may be normal in children, but there should be no associated translation at any level (56). The interspinous distance at each level should be evaluated and should be no more than 1.5 times the distance at adjacent levels; if this ratio is greater, an injury should be suspected. There are calculated norms for the interspinous distances in children, and any value greater than two standard deviations above normal is indicative of a ligamentous injury (57). The measurement of soft tissue spaces is important in evaluating any evidence of swelling or hemorrhage, which may be associated with an occult injury. The normal retropharyngeal soft tissue space should be less than 6 mm at C3 and less than 14 mm at C6. These spaces may be increased in children without an injury who are crying at the time of the x-ray, because the attachment of the pharynx to the hyoid bone results in its forward displacement with crying, producing an apparent increase in the width of these spaces. These x-rays must be taken with the patient quiet and repeated if there is any doubt.
FIGURE 18-10 A. Pseudosubluxation of C2 on C3. B. True subluxation. In flexion the posterior element of C2 should normally align itself with the posterior elements C1 and C3. The relationship of the body of C2 with the body of C3 gives the appearance of subluxation; however, the alignment of the posterior elements of C1-C3 confirms pseudosubluxation.
FIGURE 18-11 Normal relationships in the lateral cervical spine: 1, spinous processes; 2, spinolaminar line; 3, posterior vertebral body line; 4, anterior vertebral body line. (Reprinted from Copley LA, Dormans JP. Cervical spine disorders in infants and children. J Am Acad Orthop Surg 1998;6:204–214; with permission.)
Special Imaging Studies
Most cervical spine injuries in children are detected by plain x-rays (1). Most ligamentous injuries can be identified on flexion and extension views of the cervical spine in a cooperative and alert patient. The roles of computed tomography (CT) scanning and magnetic resonance imaging (MRI) continue to evolve in the evaluation of trauma patients.
When CT scanning is used in children, a few salient points should be kept in mind. First, the proportion of a child’s head to his or her body is greater than that of an adult, so care must be taken not to position the head in flexion to obtain the scan. Inadvertent flexion may potentiate any occult fracture not seen on plain films. Second, the radiation doses for CT scanning are significantly higher than for plain x-rays. CT protocols for children should be used to limit the amount of radiation that the head and neck receive during scanning of the cervical spine. While axial views are standard, coronal and sagittal reformatted images and three-dimensional reconstruction views provide improved anatomic details of the spine and can be obtained without any additional radiation to the patient (58). In patients with head injuries, CT scanning of the cervical spine can be done at the time of CT scanning of the head to reduce the number of plain films that may be required to document that there is not a neck injury (59). However, plain x-rays remain the standard for initial evaluation of the pediatric cervical spine; CT scanning as an initial imaging study is associated with an increase in radiation with no demonstrable benefit over plain films (60).
MRI has become increasingly useful in evaluating pediatric patients with suspected cervical spine injuries (Fig. 18-12), especially for ruling out ligamentous injuries in patients who cannot cooperate with flexion and extension views (61). The advantages of an early MRI are the ability to allow mobilization if no injury is present and the early detection of an unrecognized spinal fracture to allow proper treatment. MRI also is useful in

evaluating patients with SCIWORA. MR angiography has replaced standard arteriography for evaluation of the vertebral arteries in patients with upper cervical spine injuries who have suspected arterial injuries (62). MRI also remains the best imaging modality for evaluating injuries of the intervertebral disks and is especially useful to detect disk herniation in adolescent patients with facet joint injuries that may require operative reduction.
FIGURE 18-12 MRI depicts injury to the cervical cord and upper cervical spine.
FIGURE 18-13 A. Adult immobilized on a standard backboard. B. Young child on a standard backboard. The relatively large head forces the neck into a kyphotic position. (Reprinted from Herzenberg JE, Hensinger RN, Dedrick DK, et al. Emergency transport and positioning of young children who have an injury of the cervical spine: the standard backboard may be hazardous. J Bone Joint Surg [Am] 1989;71:15–22; with permission.)
The initial management of any child suspected of having a cervical spine injury starts with immobilization in the field. Extraction from an automobile or transport to the hospital may cause damage to the spinal cord in a child with an unstable cervical spine injury if care is not taken to properly immobilize the neck. The immobilization device should allow access to the patient’s oropharynx and anterior neck if intubation or tracheostomy becomes necessary. The device should allow splintage of the head and neck to the thorax to minimize further movement.
The use of backboards in pediatric trauma patients deserves special attention because of the anatomic differences between children and adults. Compared to adults, children have a disproportionately larger head with respect to the body. This anatomic relationship causes a child’s cervical spine to be placed in flexion if immobilization is done on a standard backboard. Herzenberg et al (63) reported 10 children under the age of 7 years whose cervical spines had anterior angulation or translation on x-ray when they were placed on a standard backboard. The use of a backboard with a recess so that the head can be lowered into it to obtain a neutral position of the cervical spine is one way to avoid unnecessary flexion. Another is a split-mattress technique in which the body is supported by two mattresses and the head is supported by one mattress, allowing the cervical spine to assume a neutral position. Children younger than 8 years of age should be immobilized on a backboard using one of these techniques (Figs. 18-13 and 18-14) (64).
FIGURE 18-14 A. Young child on a modified backboard that has a cutout to the recess of the occiput, obtaining better supine cervical alignment. B. Young child on modified backboard that has a double-mattress pad to raise the chest, obtaining better supine cervical alignment. (Reprinted from Herzenberg JE, Hensinger RN, Dedrick DK, et al. Emergency transport and positioning of young children who have an injury of the cervical spine: the standard backboard may be hazardous. J Bone Joint Surg [Am] 1989;71:15–22; with permission.)
Cervical collars supplement backboards for immobilization in the trauma setting. While soft collars tend to be more comfortable and cause less soft tissue irritation, rigid collars are preferred for patients with acute injuries because they provide better immobilization. Even rigid collars may allow up to 17 degrees of flexion, 19 degrees of extension, 4 degrees of rotation, and 6 degrees of lateral motion (65,66). Supplemental sandbags

and taping on either side of the head are recommended in all children and have been shown to limit the amount of spinal motion to 3 degrees in any plane (39).
Further displacement of an unstable cervical injury may occur if resuscitation is required. The placement of pediatric patients on an appropriate board with the neck in a neutral position makes recognition of some fractures difficult because positional reduction may have occurred, especially with ligamentous injuries or endplate fractures. An apparently normal lateral x-ray in a patient with altered mental status or multiple injuries does not rule out a cervical spine injury. A study of four patients with unstable cervical spine injuries who had attempted resuscitation in the emergency department showed that axial traction actually increased the deformity (32). Any manipulation of the cervical spine, even during intubation, must be done with caution and with the assumption that the patient has an unstable cervical spine injury until proven otherwise.
Immobilization of the cervical spine may continue after the emergency setting if there is an injury that requires treatment. Specific injuries and their treatment are described later in this chapter. Further immobilization of some cervical spine injuries requires a cervical collar. A rigid collar can be used for immobilization if it is an appropriately fitting device with more padding than a standard cervical collar placed in the emergency department. More unstable or significant injuries can be treated with a custom orthosis, a Minerva cast, or a halo device. An advantage of custom devices is the ability to use lightweight thermoplastic materials that can be molded better to each patient’s anatomy and can be extended to the thorax (Fig. 18-15). These devices must be properly applied for effective immobilization, and skin breakdown, especially over the chin region, needs to be carefully monitored. Minerva casts tend to provide more immobilization than thermoplastic devices, but their use is not as common and their application requires attention to detail.
FIGURE 18-15 Custom-made cervicothoracic brace used to treat a C2 fracture that reduced in extension.
A halo device can be used for the treatment of cervical spine injuries even in children as young as 1 year old. The halo can be used as either a ring alone to apply traction or with a vest for definitive immobilization of an unstable cervical spine injury. The complication rate related to the use of a halo in one series of patients was 68%; however, all patients were able to wear the halo until fracture healing occurred or arthrodesis (67). The most common complications in this series were superficial pin track infection and pin loosening. Other complications that occur less frequently include dural penetration, supraorbital nerve injury, unsightly pin scars, and deep infection (67,68). Prefabricated halo vests are used in adults and are easily fitted to older adolescents. Because of the age and size ranges of children, however, a custom vest or even a cast vest may be needed. Prefabricated vests are available in sizes for infants, toddlers, and children, with measurements based on the circumference of the chest at the xiphoid process. Improper fitting of a vest may allow unwanted movement of the neck despite the halo, and any size mismatch requires a custom vest or cast vest (Fig. 18-16).
The fabrication of a halo for any patient needs to consider both the size of the ring and the size of the vest. Prefabricated rings and prefabricated vests are available for even for the smallest of patients and are based on circumferential measurements at the crown and at the xiphoid process. If the size of the patient or the anatomy of the patient does not fit within these standard sizes, the fabrication of a custom halo may be necessary. Mubarak

et al (69) recommended the following steps in the fabrication of a custom halo for a child: (a) The size and configuration of the head are obtained with the use of the flexible lead wire placed around the head, (b) the halo ring is fabricated by constructing a ring 2 cm larger in diameter than the wire model, (c) a plaster mold of the trunk is obtained for the manufacture of a custom bivalved polypropylene vest, and (d) linear measurements are made to ensure appropriate length of the superstructure.
FIGURE 18-16 A. Custom halo vest and superstructure. B. In the multiple-pin, low-torque technique, 10 pins are used for an infant halo ring attachment. Usually, four pins are placed anteriorly, avoiding the temporal region, and the remaining six pins are placed in the occipital area. (Reprinted from Mubarak SJ, Camp JF, Fuletich W, et al. Halo application in the infant. J Pediatr Orthop 1989;9:612–613; with permission.)
The placement of pins into an immature skull deserves special attention because of the dangers of inadvertent skull penetration with a pin. CT scanning before halo application aids in determining bone structure and skull thickness. It also aids in determining whether or not cranial suture interdigitation is complete and if the fontanels are closed. The thickness of the skull varies greatly up to 6 years of age and is not similar to that of adults until the age of 16 years (47). Garfin et al (70) evaluated the pediatric cranium by CT and determined that the skull is thickest anterolaterally and posterolaterally, making these the optimal sites for pin placement.
The number of pins used for placement of a ring and the insertion torques used in younger children also deserve special mention. The placement of pins at the torque pressures used in adults will lead to penetration during insertion (47). Pins should be inserted at torques of 2 to 4 inch-pounds; however, the variability and reliability of pressures found with various torque wrenches during cadaver testing are great, and each pin must be inserted cautiously (71). The use of 8 to 12 pins inserted at lower torque pressures aids in obtaining a stable ring with less chance of inadvertent penetration (Fig. 18-17). The insertion of each pin perpendicular to the skull also improves the pin–bone interface and the overall strength of the construct (72). We have had success using halo vests even in children younger than 2 years of age by using multiple pins inserted to finger-tightness rather than relying on torque wrenches.
FIGURE 18-17 Shaded area represents the “safe zone” for pin placement, avoiding the supraorbital and supratrochlear nerves anteriorly and the temporalis posteriorly. (Reprinted from Crawford H. Traction. In: Weinstein SL, ed. Pediatric spine surgery, 2nd ed. Philadelphia: Lippincott Williams & Wilkins, 2001; with permission.)
Technique of Halo Application
A halo can be applied in older children and adolescents with a local anesthetic; however, in most younger children a general anesthetic should be used. The patient is positioned on the operating table in a position that prevents unwanted flexion of the neck and maintains the proper relationship of the head and neck with the trunk. The area of skin in the region of pin insertion is cleaned with antiseptic solution and appropriate areas are shaved as needed for pin placement posteriorly. The ring is placed while an assistant holds the patient’s head; it should be placed just below the greatest circumference of the skull, which corresponds to just above the eyebrows anteriorly and 1 cm above the tips of the earlobes laterally. We recommend injection of local anesthetic into the skin and periosteum through the ring holes in which the pins will be placed. The pins are placed with sterile technique.
To optimize pin placement, a few points should be kept in mind. The thickest area of the skull is anterolaterally and posterolaterally, and pins inserted at right angles to the bone have greater force distribution and strength (70,72). Anterior pins should be placed to avoid the anterior position of the supraorbital and supratrochlear nerves (Fig. 18-18). Placement of the anterior pins too far laterally will lead to penetration of the temporalis muscle, which can lead to pain with mastication and talking, as well as early pin loosening. The optimal position for the anterior pins is in the anterolateral skull, just above the lateral two thirds of the orbit and just below the greatest circumference of the skull. The posterior pins are best placed posterolaterally directly diagonal from the anterior pins. We also recommend placing the pins to finger-tightness originally and tightening two directly opposing pins simultaneously. During placement of the pins, meticulous attention should be paid to the position of the ring in order to have a circumferential fit on the patient’s skull and to avoid any pressure of the ring on the scalp, especially posteriorly.
The number of pins used and the torque pressures applied vary according to the age of the patient. In infants and younger children, we recommend the placement of multiple pins (8 to 12) tightened to finger-tightness or 2 to 4 inch-pounds to avoid unwanted skull penetration. In older children, six to eight pins are used and tightened to 4 inch-pounds. In general, in adolescents four to eight pins can be tightened with a standard torque wrench to 6 to 8 inch-pounds. Once the pins are tightened, they must be fastened to the ring by the appropriate lock nuts or set screws. The halo vest and superstructure are then applied, with care to maintain the position of the head and neck. Appropriate positioning of the head and neck can be done by adjusting the superstructure (see Fig. 18-18).
Daily pin care should consist of hydrogen peroxide/saline cleaning at the pin–skin interface. Retightening of pins at 48 hours should be avoided in infants and children to prevent skull penetration; however, in adolescents the pins can be retightened

at 48 hours with a standard torque wrench. Local erythema or drainage may occur about the pins and can be managed with oral antibiotics and continued pin site care. If significant loosening occurs or the infection is more serious, the pin or pins should be removed. Occasionally, a dural puncture occurs during pin insertion or during the course of treatment. This necessitates pin removal and prophylactic antibiotics until the tear heals, usually at 4 to 5 days.
FIGURE 18-18 Child immobilized in a halo for C1-C2 rotary subluxation. Note the position of the anterior pins, as well as the placement of the posterior pins at 180 degrees opposite the anterior pins.
Spinal cord injury without radiographic abnormalities (SCIWORA), a syndrome first brought to the attention of the medical community by Pang and Wilberger (9), is unique to children. This condition is defined as a spinal cord injury in a patient with no visible fracture or dislocation on plain x-rays, tomograms, or CT scans.
A complete or incomplete spinal cord lesion may be present, and the injury usually results from severe flexion or distraction of the cervical spine. SCIWORA is believed to occur because the spinal column (vertebrae and disk space) in children is more elastic than the spinal cord and can undergo considerable deformation without being disrupted (73,74). The spinal column can elongate up to 2 inches without disruption, whereas the spinal cord ruptures with only a quarter-inch of elongation.
SCIWORA also may represent an ischemic injury in some patients, although most are believed to be due to a distraction-type injury in which the spinal cord has not tolerated the degree of distraction but the bony ligamentous elements have not failed. Aufdermaur suggested another possibility (13): a fracture through a pediatric vertebral endplate reduces spontaneously (much like a Salter I fracture), giving a normal x-ray appearance, although the initial displacement could have caused spinal cord injury.
SCIWORA abnormalities are more common in children under 8 years of age than in older children (9,10,56,75), perhaps because of predisposing factors such as cervical spine hypermobility, ligamentous laxity, and an immature vascular supply to the spinal cord. The reported incidence of this condition varies from 7% to 66% of patients with cervical spine injuries (9,14,76).
Delayed onset of neurologic symptoms has been reported in as many as 52% of patients in some series (9,77). Pang and Pollock reported 15 patients who had delayed paralysis after their injuries (14). Nine had transient warning signs such as paresthesia or subjective paralysis. In all patients with delayed onset of paralysis, the spine had not been immobilized after the initial trauma, and all were neurologically normal before the second event. This underlines the importance of diligent immobilization of a suspected spinal cord injury in a child. Approximately half of the young children with SCIWORA in reported series had complete spinal cord injuries, whereas the older children usually had incomplete neurologic deficit injuries that involved the subaxial cervical spine (13,77,78,79).
Careful x-ray evaluation is helpful in the workup of these patients, but MRI will show a spinal cord lesion that often is some distance from the vertebral column injury. As many as 5% to 10% of children with spinal cord injuries have normal x-ray results (3,80).
Spinal cord injuries are still rare in children. Rang reviewed spinal injuries at the Toronto Hospital for Sick Children over 15 years and found that children constituted a small percentage of the patients with acquired quadriplegia or paraplegia (81). He found that paraplegia was three times more common than quadriplegia. When a spinal cord injury is suspected, the neurologic examination must be complete and meticulous and may

take several examinations of sensory and motor function. If an acute spinal cord injury is documented by examination, the administration of methylprednisolone within the first 8 hours after injury has been shown to improve the chances of neurologic recovery (82,83,84,85). Methylprednisolone in the treatment of acute spinal cord injuries has been shown to improve motor and sensory recovery when evaluated 6 weeks and 6 months after injury (84); however, this positive effect on neurologic recovery is limited to those treated within the first 8 hours of injury. The initial loading dose of methylprednisolone is 30 mg/kg body weight. If the loading dose is given within 3 hours after injury, then a maintenance infusion of 5.4 mg/kg is given for 24 hours after injury. If the loading dose is given between 3 and 8 hours after injury, then a maintenance infusion of 5.4 mg/kg is given for 48 hours after injury. Methylprednisolone decreases edema, has an anti-inflammatory effect, and protects the cell membranes from scavenging oxygen free radicals (82,83,84,85).
In several series (82,83,84,85) there was a slight increase in the incidence of wound infections but no significant increase in gastrointestinal bleeding. All of these studies involved patients 13 years or older, so no documentation of the efficacy in young children exists. A combination of methylprednisolone and GM1 ganglioside is being studied for its possible beneficial effect on an injured spinal cord (86,87,88,89). GM1 is a complex acid-like lipid found at high levels in the cell membrane of the central nervous system that is thought to have a neuroprotective and neurofunctional restorative potential. Early studies have shown that patients given both drugs had improved recovery over those who had received just methylprednisolone.
Once spinal cord injury is documented, routine care includes prophylaxis for stress ulcers, routine skin care to prevent pressure sores, and initial Foley catheterization followed by intermittent catheterization and a bowel training program. With incomplete lesions, children have a better chance than adults for useful recovery. Hadley et al (3) noted that 89% of pediatric patients with incomplete spinal cord lesions improved, whereas only 20% of patients with complete injuries had evidence of significant recovery. Laminectomy has not been shown to be beneficial and can actually be harmful (90,91) because it increases instability in the cervical spine; for example, it can cause a swan-neck deformity or progressive kyphotic deformity (92,93). The risk of spinal deformity after spinal cord injury has been investigated by several researchers (22,92,94,95,96,97). Mayfield et al (92) found that patients who had a spinal cord injury before their growth spurt all developed spinal deformities, 80% of which were progressive. Ninety-three percent developed scoliosis, 57% kyphosis, and 18% lordosis. Sixty-one percent of these patients required spinal arthrodesis for stabilization of their curves. Orthotic management usually is unsuccessful, but in some patients it delays the age at which arthrodesis is necessary. Lower extremity deformities also may occur, such as subluxations and dislocations about the hip. Pelvic obliquity can be a significant problem and may result in pressure sores and difficulty in seating in a wheelchair.
Spinal column injury and spinal cord injury can occur during birth, especially during a breech delivery (98,99). Injuries associated with breech delivery usually are in the lower cervical spine or upper thoracic spine and are thought to result from traction, whereas injuries associated with cephalic delivery usually occur in the upper cervical spine and are thought to result from rotation. It is unclear whether cesarean section reduces spinal injury in neonates (100); however, Bresnan and Abroms (101) noted that neck hyperextension in utero (star-gazing fetus) in breech presentations is likely to result in an estimated 25% incidence of spinal cord injury with vaginal delivery and can be prevented by delivering by cesarean section.
Distraction-type injuries to the upper cervical spine have been reported in infants in forward-facing car seats. Because infants have poor head control and muscular development, if they are placed in a forward-facing car seat and a sudden deceleration occurs, the head continues forward while the remainder of the body is strapped in the car seat, resulting in a distraction-type injury (102,103).
Neuromuscular control of the cervical spine in neonates and infants is underdeveloped, and a normal infant cannot adequately support his or her head until about 3 months of age. Infants, therefore, cannot protect their spines against excessive forces that may occur during delivery or during the months after birth. Skeletal injuries from obstetric trauma are probably underreported because the infantile spine is largely cartilaginous and difficult to evaluate with x-rays, especially if the injury is through the cartilage or cartilage–bone interface (13). A cervical spine lesion should be considered in an infant who is floppy at birth, especially after a difficult delivery. Flaccid paralysis with areflexia usually is followed by a typical pattern of hyper-reflexia once spinal cord shock is over. Brachial plexus palsies also warrant cervical spine x-rays. MRI can sometimes be helpful in this diagnosis.
Shulman et al (104) found atlanto-occipital and axial dislocations at autopsy, and Tawbin (99) found a 10% incidence of brain and spinal injuries at autopsy.
Treatment of neonatal cervical spine injuries is nonoperative and should consist of careful realignment and positioning of the child on a bed with neck support or a custom cervical thoracic orthosis. Healing of bony injuries usually is rapid and complete (11).
Caffey (28) in 1974 and Swischuk (105) in 1969 described a child abuse syndrome called the shaken infant syndrome. Children have weak and immature neck musculature and cannot support their heads when they are subjected to whiplash stresses. Intercranial and interocular hemorrhages can occur. This injury can result in death or cerebral injury with retardation and permanent visual and hearing defects. Fractures of the spinal column and spinal cord injuries can occur during violent shaking of a child. Swischuk reported a spinal cord injury in a 2-year-old that was the result of violent shaking that produced a cervical fracture dislocation that spontaneously reduced (105).
TABLE 18-2 Anderson and Montesano Classification of Occipital Condylar Fractures
Type Description Biomechanics
I Impaction Results from axial loading; ipsilateral alar ligament may be compromised, but stability is maintained by contralateral alar ligament and tectorial membrane.
II Skull base extension Extends from occipital bone via condyle to enter foramen magnum; stability is maintained by intact alar ligaments and tectorial membrane.
III Avulsion Mediated via alar ligament tension; associated disruption of tectorial membrane and contralateral alar ligament may cause instability.
(Modified from Hanson JA, Deliganis AV, Baxter AB, et al. Radiologic and clinical spectrum of occipital condyle fractures: retrospective review of 107 consecutive patients in 95 patients. AJR Am J Roentgenol 2002;178:1261–1268.)

Occipital condylar fractures are rare, and their diagnosis requires a high index of suspicion (106,107). CT with multiplanar reconstruction usually is necessary to establish the diagnosis. Tuli et al (108) recommended that a CT scan be obtained in the following circumstances: presence of lower cranial nerve deficits, associated head injury or basal skull fracture, or persistent neck pain despite normal x-rays. Reports of associated cranial nerve deficits vary from 53% to 31% of patients with occipital condylar fractures (1,79,108). Anderson and Montesano (1) described three types of occipital condylar fractures (Table 18-2, Fig. 18-19): type I, impaction fracture; type II, basilar skull fracture extending into the condyle; and type III, avulsion fractures. An avulsion fracture is the only type of occipital condylar fracture that is unstable. Type I injuries are the result of axial compression with a component of ipsilateral flexion. Type II injuries are basilar skull fractures that extend to involve the occipital condyle and usually are caused by a direct blow. Type III injuries are avulsion fractures of the inferomedial portion of the condyle that is attached to the alar ligament. Types I and II occipital condylar fractures usually are stable and can be treated with a cervical orthosis. Type III or avulsion fractures can be unstable and may require halo immobilization or occipitocervical arthrodesis (109).
TABLE 18-3 Tuli et al’s Classification of Occipital Condylar Fractures
Type Description Biomechanics
1 Nondisplaced Stable
2A Displaced* Stable; no x-ray, CT, or MRI evidence of occipitoatlantoaxial instability of ligamentous disruption
2B Displaced* Unstable; positive x-ray, CT, or MRI evidence of occipitoatlantoaxial instability or ligamentous disruption
*At least 2 mm of osseous separation.
(Modified from Hanson JA, Deliganis AV, Baxter AB, et al. Radiologic and clinical spectrum of occipital condyle fractures: retrospective review of 107 consecutive patients in 95 patients. AJR Am J Roentgenol 2002;178:1261–1268.)
Tuli et al (108) also classified occipital condylar fractures based on displacement and stability of the occiput/C1-C2 complex (Table 18-3). In their classification, type 1 fractures are nondisplaced and type 2 are displaced. They further subdivided type 2 fractures into type 2A, displaced but stable, and type

2B, displaced and unstable. Most occipital condylar fractures can be treated with a cervical orthosis, but unstable fractures may need halo immobilization or occipital cervical fusion.
FIGURE 18-19 Classification of occipital condylar fractures according to Anderson and Monsanto. A. Type I fractures can occur with axial loading. B. Type II fractures are extensions of basilar cranial fractures. C. Type III fractures can result from an avulsion of the condyle during rotation, lateral bending, or a combination of mechanisms. (Reprinted from Hadley MN. Occipital condyle fractures. Neurosurgery 2002;50[Suppl]:S114–S119; with permission.)
Atlanto-occipital dislocation was once thought to be a rare fatal injury found only at the time of autopsy (Fig. 18-20) (13,82,104,110,111). This injury is now being recognized more often, and children are surviving (67,112,113,114). This increase in the survival rate may be due to increased awareness and improved emergency care with resuscitation and spinal immobilization by emergency personnel. Atlanto-occipital dislocation occurs in sudden deceleration accidents, such as motor vehicle or pedestrian–vehicle accidents. The head is thrown forward, and this can cause sudden craniovertebral separation.
The atlanto-occipital joint is a condylar joint that has little inherent bony stability. Stability is provided by the ligaments about the joint. The primary stabilizers are the paired alar ligaments, the articular capsule, and the tectorial membrane (a continuation of the posterior longitudinal ligament). In children, this articulation is not as well formed as in adults and it is less cup-shaped. Therefore, there is less resistance to translational forces (13,52,104,110,111,115).
Diagnosis may be difficult because atlanto-occipital dislocation is a ligamentous injury. Although patients with this injury have a history of trauma, some may have no neurologic findings. Others, however, may have symptoms such as cranial nerve injury, vomiting, headache, torticollis, or motor or sensory deficits (111,113,116,117,118,119). Brain stem symptoms, such as ataxia and vertigo, may be caused by vertebrobasilar vascular insufficiency. Unexplained weakness or difficulty in weaning off a ventilator after a closed head injury may be a sign of this injury.
FIGURE 18-20 Patient with atlanto-occipital dislocation. Note the forward displacement of the Wackenheim line and the significant anterior soft tissue swelling.
The treating physician must have a high index of suspicion in children with closed head injuries or associated facial trauma and must be aware of the x-ray findings associated with atlanto-occipital dislocation. A significant amount of anterior soft tissue swelling usually can be seen on a lateral cervical spine x-ray. This increased anterior soft tissue swelling should be a warning sign that an atlanto-occipital dislocation may have occurred.
X-ray findings that aid in the diagnosis of atlanto-occipital dislocation are the Wackenheim line, Powers ratio, dens–basion interval, and occipital condylar distance. The Wackenheim line is drawn along the clivus and should intersect tangentially the tip of the odontoid. A shift anterior or posterior of this line represents either an anterior or posterior displacement of the occiput on the atlas (Fig. 18-21). This line is probably the most helpful because it is reproducible and easy to identify on a lateral x-ray. The Powers ratio (see Fig. 18-7) is determined by drawing a line from the basion to the posterior arch of the atlas (BC) and a second line from the opisthion to the anterior arch of the atlas (OA). The length of line BC is divided by the length of the line OA, producing the Powers ratio. A ratio of more than 1.0 is diagnostic of anterior atlanto-occipital dislocation. A ratio of less than 0.7 is diagnostic of posterior atlanto-occipital dislocation. Values between 1.0 and 0.7 are considered normal. Another x-ray measurement is the dens–basion interval. If the interval measures more than 1.2 cm, then disruption of the atlanto-occipital joint has occurred (52,120). Donahue et al (51) described an occipital condylar facet distance of more than 5 mm from the occipital condyle to the C1 facet as indicative of atlanto-occipital injury. They recommended measuring this distance from five reference points along the occipital condyle and the C1 facet (Fig. 18-22).
MRI is useful in diagnosing atlanto-occipital dislocation by showing soft tissue edema around the tectorial membranes and lateral masses and ligament injury or disruption (121).
Operative Treatment
Because atlanto-occipital dislocation is a ligamentous injury, nonoperative treatment usually is unsuccessful. Although Farley et al (122) reported successful stabilization in a halo, Georgopoulos et al (123) found persistent atlanto-occipital instability after halo immobilization. Immobilization in a halo should be used with caution: if the vest or cast portion is not fitted properly, displacement can increase (Fig. 18-23) because the head is fixed in the halo but movement occurs because of inadequate immobilization of the trunk in the brace or cast. Traction should be avoided because it can cause distraction of the skull from the atlas. Surgical stabilization is the recommended treatment (124). Posterior arthrodesis can be performed in situ, with wire fixation or fixation with a contoured Luque rod and wires (125,126,127). If the C1-C2 articulation is stable, arthrodesis should be only from the occiput to C1 so that C1-C2 motion is preserved (128). Some researchers have expressed reservations about the chance of obtaining fusion in the narrow atlanto-occipital interval and have


recommended arthrodesis from the occiput to C2. If stability of the C1-C2 articulation is questionable, arthrodesis should extend to C2 (129). Acute hydrocephalus can occur after this injury or in the early postoperative period because of changes in cerebrospinal fluid flow at the cranial cervical junction.
FIGURE 18-21 Craniovertebral dislocation. A. Lateral view shows extensive soft tissue swelling. The distance between the basion and the dens is 2.4 cm (arrows) (normal is <1 cm). B. Line drawing shows the abnormal relationship between the occiput and the upper cervical spine. (Reprinted from El-Khoury GY, Kathol MH. Radiographic evaluation of cervical trauma. Semin Spine Surg 1991;S3:3–23; with permission.)
FIGURE 18-22 Atlanto-occipital joint measurement points 1 through 5 demonstrated on a normal cross-table lateral skull x-ray in an 8-year-old (A) and a 14-year-old (B). (Reprinted from Kaufman RA, Carroll CD, Buncher CR. Atlantooccipital junction: standards for measurement in normal children. AJNR Am J Neuroradiol 1987;S8:995–999; with permission.)
FIGURE 18-23 A. Lateral x-ray of a patient with atlanto-occipital dislocation. Note the increase in the facet condylar distance. B. Lateral x-ray after occipital C1 arthrodesis.
For a patient who presents very late with an unreduced dislocation, an in situ arthrodesis is recommended. DiBenedetto and Lee (130) recommended arthrodesis in situ with a suboccipital craniectomy to relieve posterior impingement.
Instability at the atlanto-occipital joint is increased in patients with Down syndrome as well as in those with a high cervical arthrodesis below the axis. These patients may be at risk of developing chronic instability patterns and are at higher risk of having instability after trauma.
FIGURE 18-24 Technique of occipitocervical arthrodesis used when the posterior arch of C1 is absent. A. Exposure of the occiput, atlas, and axis. B. Reflection of periosteal flap to cover defect in atlas. C. Decortication of exposed vertebral elements. D. Placement of autogenous cancellous iliac bone grafts. (Redrawn from Koop SE, Winter RB, Lonstein JE. The surgical treatment of instability of the upper part of the cervical spine in children and adolescents. J Bone Joint Surg [Am] 1984;66:403; with permission.)
Occiput to C2 Arthrodesis
Arthrodesis Without Internal Fixation
In younger children in whom the posterior elements are absent at C1 or separation is extensive in the bifid part of C1 posteriorly, posterior cervical arthrodesis from the occiput to C2 with iliac crest bone graft is performed using a periosteal flap from the occiput to provide an osteogenic tissue layer for the bone graft (Fig. 18-24) (131).
A halo is applied after the patient is anesthetized, endotracheal intubation is obtained, and all anesthesia lines are in place. For younger children, 8 to 12 pins with lower-pressure torque are used in the halo (see Fig. 18-17); in older children, 4 pins can be used.
An x-ray is obtained to evaluate the position of the head and cervical spine in the prone position with the halo in place. The

x-ray also aids in identifying landmarks and levels, although once the skin incision is made, the occiput and spinous processes can be palpated.
A straight posterior incision is made from the occiput to about C3, with care not to expose below C2 to avoid extension of the fusion to lower levels. An epinephrine and lidocaine solution is injected into the cutaneous and subcutaneous tissues to help control local skin and subcutaneous bleeding. The incision is deepened in the midline to the spinous processes of C2. Once identified, the level of the posterior elements of C1 or the dura is more easily found.
After C2 is identified, subperiosteal dissection is carried proximally. Extraperiosteal dissection is used to approach the occiput (see Fig. 18-24A). The dura is not completely exposed; if possible, any fat or ligamentous tissue present is left intact. The interspinous ligaments also should be left intact.
The occipital periosteum is mobilized by making a triangular incision directly on the posterior skull, with the apex posteriorly and the broad base over the foramen magnum region. A flap of 3 or 4 cm at the base can be created. With subperiosteal elevation, the periosteum can be reflected from the occiput to the spinous processes of C2 (see Fig. 18-24B). The apex of the flap is sutured to the spinous process of C2 and is attached laterally to any posterior elements that are present at C1 or other lateral soft tissues. After the periosteum is secured to the bone and any rudimentary C1 ring is exposed subperiosteally, a power bur is used to decorticate the occiput and any exposed portions of C1 and C2 (see Fig. 18-24C).
FIGURE 18-25 Technique of occipitocervical arthrodesis used in older adolescents with intact posterior elements of C1 and C2. A. A bur is used to create a ridge in the external occipital protuberance, and then a hole is made in the ridge. B. Wires are passed through the outer table of the occiput, under the arch of the atlas, and through the spinous process of the axis. C. Corticocancellous bone grafts are placed on the wires. D. Wires are tightened to secure grafts in place. (Redrawn from Wertheim SB, Bohlman HH. Occipitocervical fusion: indications, technique, and long-term results. J Bone Joint Surg [Am] 1987;69:833; with permission.)
Iliac crest bone graft is harvested, and struts of iliac bone are placed across the area on the periosteal flap (see Fig. 18-24D). No internal fixation is used other than sutures to secure the periosteum. The wound is closed in a routine fashion, and a body jacket or cast is applied and attached to the halo. The halo cast is worn until x-rays show adequate posterior arthrodesis, usually in 8 to 12 weeks.
Arthrodesis with Triple-Wire Fixation
In older adolescents in whom the posterior elements of C1 and C2 are intact, a triple-wire technique, as described by Wertheim and Bohlman (132), can be used (Fig. 18-25). The wires are passed through the

outer table of the skull at the occipital protuberance. Because the transverse and superior sagittal sinuses are cephalad to the protuberance, they are not endangered by wire passage.
Stability of the spine is obtained preoperatively with cranial skeletal traction with the patient on a turning frame or cerebellar head rest. The patient is placed prone, and a lateral x-ray is obtained to document proper alignment. The subcutaneous tissues are injected with an epinephrine solution (1:500,000). A midline incision is made extending from the external occipital protuberance to the spine of the third cervical vertebra. The paraspinous muscles are sharply dissected subperiosteally with a scalpel, and a periosteal elevator is used to expose the occiput and cervical laminae, with special care to stay in the midline to avoid the paramedian venous plexus. At a point 2 cm above the rim of the foramen magnum, a high-speed diamond bur is used to create a trough on either side of the protuberance, making a ridge in the center (see Fig. 18-25A). A towel clip is used to make a hole in this ridge through only the outer table of bone. A 20-gauge wire is looped through the hole and around the ridge; then another 20-gauge wire is looped around the arch of the atlas. A third wire is passed through a hole drilled in the base of the spinous process of the axis and around this structure, giving three separate wires to secure the bone grafts on each side of the spine (see Fig. 18-25B).
A thick, slightly curved graft of corticocancellous bone of premeasured length and width is removed from the posterior iliac crest. The graft is divided horizontally into two pieces, and three holes are drilled into each graft (see Fig. 18-25C). The occiput is decorticated and the grafts are anchored in place with the wires on both sides of the spine (see Fig. 18-25D). Additional cancellous bone is packed around and between the two grafts. The wound is closed in layers over suction drains.
Either a rigid cervical orthosis or a halo cast is worn for 6 to 15 weeks, followed by a soft collar that is worn for an additional 6 weeks.
Occipitocervical Arthrodesis
The positioning of the patient and the procedure are done with the patient under general anesthesia and with monitoring of the somatosensory evoked potentials (Fig. 18-26). A halo ring is applied initially with the patient supine. Subsequently, the patient is carefully placed in the prone position, the halo is secured to the operating table with a halo-positioning device, and the alignment of the occiput and the cervical spine is confirmed with a lateral x-ray. The midline is exposed from the occiput to the second or third cervical vertebra. Particular care is taken to limit the lateral dissection to avoid damaging the vertebral arteries (133).
In patients who need decompression because of cervical stenosis or for removal of a tumor, the arch of the first or second cervical vertebra (or both) is removed, with or without removal of a portion of the occipital bone to enlarge the foramen magnum.
Four holes, aligned transversely, with two on each side of the midline, are made with a high-speed drill through both cortices of the occiput, leaving a 1-cm osseous bridge between the two holes of each pair. The holes are placed caudal to the transverse sinuses. A trough is fashioned into the base of the occiput to accept the cephalad end of the bone graft. A corticocancellous graft is obtained from the iliac crest and is shaped into a rectangle, with a notch created in the inferior base to fit around the spinous process of the second or third cervical vertebra. The caudal extent of the intended arthrodesis (the second or third cervical vertebra) is determined by the presence or absence of a previous laminectomy, congenital anomalies, or

the level of the instability. On each side, a looped 16- or 18-gauge Luque wire is passed through the bur holes and looped on itself. Wisconsin button wires (Zimmer, Warsaw, IN) are passed through the base of the spinous process of either the second or the third cervical vertebra. The wire that is going into the left arm of the graft is passed through the spinous process from right to left. The graft is placed into the occipital trough superiorly and about the spinous process of the vertebra that is to be at the caudal level of the arthrodesis (the second or third cervical vertebrae). The graft is precisely contoured so that it fits securely into the occipital trough and around the inferior spinous process before the wires are tightened. The wires are subsequently crossed, twisted, and cut. An intraoperative x-ray is made at this point to assess the position of the graft and the wires as well as the alignment of the occiput and the cephalad cervical vertebrae. Extension of the cervical spine can be controlled by positioning of the head with the halo frame, by adjustment of the size and shape of the graft, and to a lesser extent by appropriate tightening of the wires.
FIGURE 18-26 Occipitocervical arthrodesis. A. Four bur holes are placed into the occiput in transverse alignment, with two on each side of the midline, leaving a 1-cm osseous bridge between the two holes of each pair. A trough is fashioned into the base of the occiput. B. 16- or 18-gauge Luque wires are passed through the bur holes and looped on themselves. Wisconsin button wires are passed through the base of the spinous process of either the second or third cervical vertebra. The graft is positioned into the occipital trough and spinous process of the cervical vertebra at the caudal extent of the arthrodesis. The graft is locked into place by the precise contouring of the bone. C. The wires are crossed, twisted, and cut. The extension of the cervical spine can be controlled by positioning of the head with the halo frame, by adjustment of the size and shape of the bone graft, and to a lesser extent by tightening of the wires. (Reprinted from Dormans JP, Drummond DS, Sutton LN, et al. Occipitocervical arthrodesis in children. J Bone Joint Surg [Am] 1995;77:1234–1240; with permission.)
Atlanto-Occipital Arthrodesis
Although most patients with atlanto-occipital dislocations are treated with fusion from the occiput to C2 or lower, Sponseller and Cass (128) described occiput–C1 fusion in two children with atlanto-occipital arthrodesis who had complete or near-complete neurologic preservation. Their rationale was that rotation would be preserved by sparing the C1-C2 articulation from fusion and that less stress would be concentrated on the lower cervical spine by fusing one level instead of two. In both of their patients, stable fusion was obtained and neurologic status was maintained.
Before surgery, x-rays and CT scans should be reviewed to be sure a bifid or hypoplastic C1 arch is not present. The positioning of the patient and the procedure are done with the patient under general anesthesia and with monitoring of the somatosensory evoked potentials. The procedure is done with the patient immobilized in a halo vest. The base of the skull to the ring of C1 is exposed and the periosteum of the skull is elevated so that it forms a flap from the foramen magnum located posteriorly-superiorly. The ring of C1 is carefully exposed, with care taken not to dissect more than 1 cm to either side of the midline to protect the vertebral arteries. Care also is taken not to expose any portion of C2 to prevent bridging of the fusion. The dissection of C1 should be done gently. A trough for the iliac crest bone graft is made in the occiput at a level directly cranial to the ring of C1. This trough is unicortical only and extends the width of the exposed portion of C1. Superior to this, two holes are drilled through the occiput as close to the trough as possible to avoid an anteriorly translating vector on the skull when tightening it down to C1. One 22-gauge wire is passed through the holes and another is placed around the ring of C1. The periosteal flap is turned down to bridge the occiput–C1 interval. A small, rectangular, bicortical, iliac crest bone graft approximately 1.5 cm wide and 1 cm high is shaped to fit the trough in the occiput; the graft is contoured to fit the individual patient’s occiput–C1 interval. The inferior surface of the bone graft is contoured to fit snugly around the ring of C1 to keep it from migrating anteriorly into the epidural space. Two holes are drilled directly above the distal end of the graft, and the wire around C1 is passed through these holes, forming two distal strands; the wire passed through the occiput forms two proximal strands. These are twisted together and sequentially tightened to apply slight compression to the bone graft. This keeps the graft in the occipital trough and prevents migration into the canal by the occiput. Additional cancellous bone is added to any available space.
The halo vest is kept in place for 6 to 8 weeks in a young child and for as long as 12 weeks in an older child or adolescent. Union is confirmed by a coned, lateral x-ray of the posterior occiput–C1 interval and by flexion-extension lateral views. A rigid cervical collar is used for an additional 2 to 4 weeks to protect the fusion and support the patient’s cervical muscles while motion is regained.
Occipitocervical Arthrodesis with Contoured Rod and Segmental Wire
Occipitocervical arthrodesis using a contoured rod and segmental wire has the advantage of achieving immediate stability of the occipitocervical junction (Fig. 18-27), which allows the patient to be immobilized in a cervical collar after surgery, avoiding the need for halo immobilization.
The base of the occiput and the spinous processes of the upper cervical vertebrae are approached through a longitudinal midline incision, which extends deep within the relatively avascular intermuscular septum. The entire field is exposed subperiosteally. A template of the intended shape of the stainless steel rod is made with the appropriate length of Luque wire. Two bur holes are made on each side, about 2 cm lateral to the midline and 2.5 cm above the foramen magnum. Care should be taken to avoid the transverse and sigmoid sinus when making these bur holes. At least 10 mm of intact cortical bone should be left between the bur holes to ensure solid fixation. Luque wires or Songer cables are passed in an extradural plane through the two bur holes on each side of the midline. The wires or cables are passed sublaminar in the upper cervical spine. The rod is bent to match the template; this usually will have a head–neck angle of about 135 degrees and slight cervical lordosis. A Bend Meister (Sofamor/Danek, Memphis, TN) may be helpful in bending the rod. The wires or cables are secured to the rod. The spine and occiput are decorticated and autogenous cancellous bone grafting is performed.
Fracture of the ring of C1 (Jefferson fracture) is caused by an axial load applied to the head and is not a common injury in children (129,134,135,136,137,138,139). The force is transmitted through the occipital condyles to the lateral masses of C1, causing a disruption in the ring of C1, usually in two places, with fractures

occurring in both the anterior and posterior rings. In children an isolated single fracture of the ring can occur with the remaining fracture hinging on a synchondrosis. This is an important distinction in children because often fractures occur through a normal synchondrosis and there can be plastic deformation of the ring. This distinction can be seen on plain x-rays, with fractures appearing through what appears to be normal physes. As the lateral masses separate, the transverse ligament may be ruptured or avulsed, resulting in C1 and C2 instability (137). If the two lateral masses are widened more than 7 mm beyond the borders of the axis on an anteroposterior x-ray, then an injury to the transverse ligament is presumed. Injury to the

transverse ligament may be from a rupture of the ligament or an avulsion of the ligament attachment to C1. Jefferson fractures may be evident on plain x-rays, but CT scans are superior at showing this injury. CT scans also can be used to follow the progress of healing.
FIGURE 18-27 Occipitocervical arthrodesis using a contoured rod and segmental wire or cable fixation. A and B reprinted from Warner WC. Pediatric cervical spine. In: Canale ST, ed. Campbell’s operative ortho-paedics. St. Louis: Mosby Year Book, 1998; with permission.)
FIGURE 18-28 X-ray of an atlas fracture.
Treatment consists of immobilization in an orthosis (rigid collar or sternal occipital mandibular immobilizer |SOMI|), Minerva cast, or halo brace. If there is excessive widening (>7 mm), halo traction followed by halo brace or cast immobilization is recommended. Surgery rarely is necessary to stabilize these fractures (Fig. 18-28).
Odontoid Fractures
Odontoid fractures are one of the most common fractures of the cervical spine in children (133), occurring at an average age of 4 years (140,141,142). The unique feature of odontoid fractures in children is that the fracture most commonly occurs through the synchondrosis of C2 distally at the base of the odontoid. This synchondrosis is a cartilage line at the base of the odontoid and looks like a physeal or Salter I injury.
A fracture of the odontoid usually is associated with head trauma from a motor vehicle accident or a fall from a height, although it also can occur after trivial head trauma. X-rays should be obtained in any child complaining of neck pain. Clinically, children with odontoid fractures complain of neck pain and resist attempts to extend the neck. Odent et al (143) reported that 8 of 15 odontoid fractures in children were the result of motor vehicle accidents, with the child fastened in a forward-facing seat. The sudden deceleration of the body as it is strapped into the car seat while the head continues to travel forward causes this fracture.
Most odontoid injuries are anteriorly displaced and usually have an intact anterior periosteal sleeve that provides some stability to the fracture when immobilized in extension and allows excellent healing of the fracture (144,145,146,147). Growth disturbances are uncommon after this type of fracture. This synchondrosis normally closes at about 3 to 6 years of age and adds little to the longitudinal growth of C2.
Most often the diagnosis can be ascertained by viewing the plain x-rays. Anteroposterior views usually appear normal, and the diagnosis must be made from lateral views because displacement of the odontoid usually occurs anteriorly. Plain x-rays sometimes can be misleading when the fracture occurs through the synchondrosis and has spontaneously reduced. When this occurs the fracture has the appearance of a nondisplaced Salter I fracture. CT scans with three-dimensional reconstruction views may be needed to fully delineate the injury (148). MRI also may be useful in nondisplaced fractures by detecting edema around the injured area, indicating that a fracture may have occurred. Dynamic flexion and extension views to demonstrate instability may be obtained in a cooperative child if a nondisplaced fracture is suspected. These studies should be done only in a cooperative child and under the direct supervision of the treating physician.
Odontoid fractures in children generally heal uneventfully and rarely have complications. Neurologic deficits rarely have been reported after this injury (143,149). Odent et al (143) described neurologic injuries in 8 of 15 patients, although most were stretch injuries to the spinal cord at the cervical thoracic junction and not at the level of the odontoid fracture.
Treatment of odontoid fractures is by closed reduction (usually extension or slight hyperextension of the neck), although complete reduction of the translation is not necessary. At least 50% apposition should be obtained to provide adequate cervical alignment, and then the patient should be immobilized in a Minerva or halo cast or custom orthosis. This fracture will heal in about 6 to 8 weeks. After bony healing, stability should be documented by flexion/extension lateral radiographs. Once the Minerva cast or halo is removed, a soft collar is worn for 1 to 2 weeks. If an adequate reduction cannot be obtained by recumbency and hyperextension, then a head halter or halo traction is needed. Rarely, manipulation under general anesthesia is needed for irreducible fractures (Fig. 18-29). Surgery with internal fixation rarely has been reported due to the good results that are achieved with conservative treatment in children (49,150,151,152).
Os Odontoideum
Os odontoideum consists of a round ossicle that is separated from the axis by a transverse gap, which leaves the apical segment without support. Fielding et al (153,154,155,156,157) suggested that this was an unrecognized fracture at the base of the odontoid. Some studies have documented normal x-rays of the dens with abnormal x-rays after trivial trauma. This can be explained by a distraction force being applied by the alar ligaments, which

pulls the tip of the fractured odontoid away from the base and produces a nonunion (158,159,160,161,162,163,164). Other authors believe this to be of congenital origin because of its association with other congenital anomalies and syndromes (165,166,167).
FIGURE 18-29 Lateral x-ray and CT reconstruction view of odontoid fracture through the synchondrosis of C2. Note the anterior displacement.
The presentation of an os odontoideum can be variable. Signs and symptoms can range from a minor to a frank compressive myelopathy or vertebral artery compression. Presenting symptoms may be neck pain, torticollis, or headaches caused by local irritation of the atlantoaxial joint. Neurologic symptoms can be transient or episodic after trauma to complete myelopathy caused by cord compression (168). Symptoms may consist of weakness and loss of balance with upper motor neuron signs, although upper motor neuron signs may be completely absent. Proprioceptive and sphincter dysfunctions also are common.
Os odontoideum usually can be diagnosed on routine cervical spine x-rays, which include an open-mouth odontoid view (Fig. 18-30). Lateral flexion and extension views should be obtained to determine if any instability is present. With os odontoideum there is a space between the body of the axis and a bony ossicle. The free ossicle of the os odontoideum usually is half the size of a normal odontoid and is oval or round, with smooth sclerotic borders. The space differs from that of an acute fracture, in which the space is thin and irregular instead of wide and smooth. The amount of instability should be documented on lateral flexion and extension plain x-rays that allow measurement of both the anterior and posterior displacement of the atlas on the axis. Because the ossicle is fixed to the anterior arch of C1 and moves with the anterior arch of C1 both in flexion and extension, measurement of the relationship of C1 to the free ossicle is of little value because they move as a unit. A more meaningful measurement is made by projecting lines superiorly from the body of the axis to a line projected inferiorly from the posterior border of the anterior arch of the atlas. This gives more information as to the stability of C1-C2. Another measurement that is very helpful is space available for the cord, which

is the distance from the back of the dens to the anterior border of the posterior arch of C1.
FIGURE 18-30 Lateral x-ray (A) and open-mouth odontoid x-ray (B) showing os odontoideum. (Reprinted from Warner WC. Pediatric cervical spine. In: Canale ST, ed. Campbell’s operative orthopaedics. St. Louis: Mosby Year Book, 1999:2817; with permission.)
Recommended treatment is posterior arthrodesis of C1 to C2. Before arthrodesis is attempted, the integrity of the arch of C1 must be documented by CT scan. Incomplete development of the posterior arch of C1 is uncommon but has been reported to occur with increased frequency in patients with os odontoideum. This may necessitate an occiput to C2 arthrodesis for stability. If a C1-C2 arthrodesis is done, one must be careful not to overreduce the odontoid and cause posterior translation. Care also must be taken in positioning the neck at the time of arthrodesis and when tightening the wires if a Gallie or Brooks arthrodesis is performed to prevent posterior translation (Figs. 18-31 and 18-32). Brockmeyer et al (169) and Wang et al (170) both reported good results with transarticular screw fixation and fusion in the treatment of children with os odontoideum. Wang et al reported the use of this technique in children as young as 3 years of age.
Traumatic Ligamentous Disruption
The transverse ligament is the primary stabilizer of an intact odontoid against forward displacement. Secondary stabilizers consist of the apical and alar ligaments, which arise from the tip of the odontoid and pass to the base of the skull. These also stabilize the atlanto-occipital joint indirectly (133). The normal distance from the anterior cortex of the dens to the posterior cortex of the anterior ring of C1 is 3 mm in adults and 4.5 mm in children. In children if the distance is more than 4.5 mm, disruption of the transverse ligament is presumed. The spinal canal at C1 is large compared with other cervical segments and accommodates a large degree of rotation and some degree of pathologic displacement without compromising the spinal cord. Steel expressed this as a rule of thirds: the spinal canal at C1

is occupied equally by the spinal cord, odontoid, and a free space, which provides a buffer zone to prevent neurologic injury. Steel found that anterior displacement of the atlas that exceeds a distance equal to the width of the odontoid may place the spinal cord at risk (171).
FIGURE 18-31 Posterior translation of atlas after C1-C2 posterior arthrodesis.
FIGURE 18-32 A. Lateral x-ray of traumatic C1-C2 instability. B. Note the increase in the atlanto–dens interval. C. Lateral x-ray after C1-C2 posterior arthrodesis.
Acute rupture of the transverse ligament is rare and reportedly occurs in fewer than 10% of pediatric cervical spine injuries (12,172). However, avulsion of the attachment of the transverse ligament to C1 may occur instead of rupture of the transverse ligament.

A patient with disruption of the transverse ligament usually has a history of cervical spine trauma and complains of neck pain, often with notable muscle spasms. Diagnosis is confirmed on lateral x-rays that show an increased atlanto–dens interval. An active flexion view may be required to show instability in cooperative patients with unexplained neck pain or neurologic findings. CT scans are useful to demonstrate avulsion of the transverse ligament from its origins to the bony ring of C1.
Although rarely used, conservative treatment of acute transverse ligament injuries has been reported. For acute injuries, reduction in extension is recommended, followed by surgical stabilization of C1 and C2 and immobilization for 8 to 12 weeks in a Minerva cast, a halo brace, or a cervical orthosis. Flexion and extension views should be obtained after stabilization to document stability.
Atlantoaxial Arthrodesis
Technique of Brooks and Jenkins
The supine patient is intubated in the supine position while still on a stretcher and is then placed prone on the operating table, with the head supported by traction; the head–thorax relationship is maintained at all times during turning (16) (Fig. 18-33). A lateral cervical spine x-ray is obtained to ensure proper alignment before surgery. The skin is prepared and draped in a sterile fashion and a solution of epinephrine (1:500,000) is injected intradermally to aid hemostasis.
FIGURE 18-33 Technique of atlantoaxial arthrodesis (Brooks-Jenkins). A. Wires are inserted under the atlas and axis. B. Full-thickness bone grafts from the iliac crest are placed between the arch of the atlas and the lamina of the axis. C, D. The wires are tightened over the graft and twisted on each side. (Redrawn from Brooks AL, Jenkins EB. Atlantoaxial arthrodesis by the wedge compression method. J Bone Joint Surg [Am] 1978;60:279; with permission.)
C1 and C2 are exposed through a midline incision. With an aneurysm needle, a Mersiline suture is passed from cephalad to caudad on each side of the midline under the arch of the atlas and then beneath the lamina of C2. These serve as guides to introduce two doubled 20-gauge wires. The size of the wire used varies depending on the size and age of the patient. Two full-thickness bone grafts approximately 1.25 × 3.5 cm are harvested from the iliac crest and beveled so that the apex of the graft fits in the interval between the arch of the atlas and the lamina of the axis. Notches are fashioned in the upper and lower cortical surfaces to hold the circumferential wires and prevent them from slipping. The doubled wires are tightened over the graft and twisted on each side. The wound is irrigated and closed in layers over suction drains.
Technique of Gallie
The supine patient is intubated while on a stretcher (173) (Fig. 18-34). The prone patient then is placed on the operating table with the head supported by traction, maintaining the head–thorax relationship during turning. A lateral cervical spine x-ray is obtained to ensure proper alignment before surgery. The skin is prepared and draped in a sterile fashion, and a solution of epinephrine (1:500,000) is injected intradermally to aid hemostasis.
A midline incision is made from the lower occiput to the level of the lower end of the fusion, extending deeply within the relatively avascular midline structures, the intermuscular septum, or ligamentum nuchae. Care should be taken not to expose any more than the area to be fused to decrease the chance of spontaneous extension of the fusion. By subperiosteal dissection, the posterior arch of the atlas and the lamina of C2 are exposed. The muscular and ligamentous attachments from C2 are removed with a curet. Care should be taken to dissect laterally along the atlas to prevent injury to the vertebral arteries and vertebral venous plexus that lie on the superior aspect of the ring of C1, less than 2 cm lateral to the midline. The upper

surface of C1 is exposed no farther laterally than 1.5 cm from the midline in adults and 1 cm in children. Decortication of C1 and C2 generally is not necessary. From below, a wire loop of appropriate size is passed upward under the arch of the atlas either directly or with the aid of a Mersiline suture. The Mersiline suture can be passed with an aneurysm needle. The free ends of the wire are passed through the loop, grasping the arch of C1 in the loop.
FIGURE 18-34 Wires are passed under the lamina of the atlas and through the spine of the axis and tied over the graft. This method is used most frequently. (Reprinted from Fielding JW, Hawkins RJ, Ratzan SA. Spine fusion for atlanto-axial instability. J Bone Joint Surg [Am] 1976;58:400; with permission.)
A corticocancellous graft is taken from the iliac crest and placed against the lamina of C2 and the arch of C1 beneath the wire. One end of the wire is passed through the spinous process of C2, and the wire is twisted on itself to secure the graft in place. The wound is irrigated and closed in layers with suction drainage tubes.
Posterior C1-C2 Transarticular Screw Fixation
Preoperative evaluation must include flexion and extension x-rays to determine whether reduction is possible (127). CT of the craniocervical region with parasagittal reconstructions through the region of the axis pars interarticularis is necessary to determine the position of the vertebral artery; an aberrant position makes the procedure impossible.
The patient is positioned prone in traction with mild neck flexion. The foramen magnum and the upper four cervical vertebrae are exposed beyond the lateral facets, and the skin is prepared to the upper thoracic region to allow for the angle of drilling. The spinous processes and the lamina and facets from C1 to C3 are exposed. The ligamentum flavum is removed between C1 and C2 and the C2 nerve is exposed and elevated. The facet joint between C1 and C2 is identified. Venous bleeding from the venous plexus around C2 is controlled with judicious bipolar cauterization and packing. The C2 pars interarticularis is identified, and satisfactory alignment of the atlas and axis is confirmed.
A bone awl or a drill is used to penetrate the posterior cortical bone of C2 at the inferior facet 3 to 4 mm from the medial edge of the facet joint. The trajectory is marked out toward the dorsal cortical aspect of the anterior arch of C1. The drill is placed 10 degrees to the vertical, pointing medially to come through the C2 pars interarticularis. In the sagittal orientation, the trajectory is slightly medial to the vertical. The length of the screw and the screw path should be measured with CT reconstructions before the procedure. Guide wire insertion is monitored with fluoroscopy, and the guide wire must pass through the C1-C2 facet joint. As the wire traverses the joint space, the atlantoaxial articulation becomes rigid and fixed. A new stiffness can be felt in the guide wire and the tap that follows. If angulation is not possible for guide wire placement with the extent of exposure, a percutaneous technique is used.
The cannulated drill is passed over the guide wire and advanced with a pneumatic drill. The guide wire is secured beyond the pneumatic drill to prevent advancement into the parapharyngeal soft tissues. A guide wire of the same length as the screw length determined with preoperative CT is passed through the cannula, and the difference between the primary guide wire and the second guide wire is measured to confirm ideal screw length.
FIGURE 18-35 Position of vertebral arteries and position of screws across atlantoaxial joint. (Reprinted from Menezes AH. Surgical approaches to the craniocervical junction. In: Weinstein SL, ed. Pediatric spine surgery, 2nd ed. Philadelphia: Lippincott Williams & Wilkins, 2001.)
Opposite-side drilling is done before screw placement so that each guide wire can be identified with lateral fluoroscopy. When both guide wires are in satisfactory position, each screw is advanced over the guide wire and its purchase obtained into the lateral mass of the atlas (Fig. 18-35). The screw head should be flush against the bone, but the screws should not be overtightened to avoid shearing of the cortex of the pars interarticularis and facet.
Interlaminar bone fusion must be done after placement of the bilateral or unilateral screw. If the vertebral artery or vein has been injured, drilling must not be done on the opposite side. Because injury to one vertebral artery probably will not cause neurologic deficit, it is important not to proceed on the opposite side if injury has occurred.
Posterior C1-C2 Transarticular Screw Fixation
Preoperative CT scans are reviewed to confirm that the vertebral artery is not in an aberrant position and that the C1-C2 anatomy is suitable for transarticular screw placement (170). The patient is placed prone with the head held in a Mayfield skull clamp. The neck is manually flexed under fluoroscopic guidance to realign the atlas and axis. A routine posterior midline incision is used to expose the posterior aspect of C1 and C2. The C2 inferior facet is used as the landmark for screw entry: the entry point is 2 mm lateral to the medial edge and 2 mm above the inferior border of the C2 facet. The drill trajectory is angled medially 5 to 10 degrees. On the lateral fluoroscopic x-ray, the drill trajectory is adjusted toward the posterior cortex of the anterior arch of C1. After tapping, a 3.5-mm lag screw is placed across the C1-C2 joint. Another screw is then placed in exactly the same way on the other side. A bone graft is harvested from either the iliac crest or occipital bone. Multistranded titanium cable is passed under the lamina of C1. The bone graft is contoured and placed between the lamina of C1 and C2 and the

wire looped around the spinal process of C2 is tightened and fixed. Care is taken not to overtighten the cable to avoid its cutting through the C1 lamina. The C1-C2 joint can be roughened with a curet and cancellous bone placed to enhance fusion. The fusion surfaces on the C1 and C2 lamina should be decorticated with a drill before the bone graft is placed.
Patients are immobilized in a hard cervical collar only; no halo or Minerva cast is used postoperatively.
Atlantoaxial Instability Associated with Congenital Anomalies and Syndromes
Although acute atlantoaxial instability in children is rare, chronic atlantoaxial instability occurs in certain conditions such as juvenile rheumatoid arthritis, Reiter syndrome, Down syndrome, and Larsen syndrome. Bone dysplasia=msuch as Morquio polysaccharidosis, spondyloepiphyseal dysplasia, and Kniest syndrome=malso may be associated with atlantoaxial instability, as well as os odontoideum, Klippel-Feil syndrome, and occipitalization of the atlas (80,174,175,176,177,178,179).
Certain cranial facial malformations have high incidences of associated anomalies of the cervical spine, such as Apert syndrome, hemifacial microsomy, and Goldenhar syndrome (180). Treatment recommendations are individualized based on the natural history of the disorder and future risk to the patient. Although there is little literature on cervical spine instability in each of these syndromes, there has been considerable interest in the incidence and treatment of atlantoaxial instability in children with Down syndrome (181,182,183,184,185,186).
Some Down syndrome patients have C1-C2 instability of more than 5 mm. The Committee on Sports Medicine of the American Academy of Pediatrics issued a policy statement in 1984 (181) asserting that Down syndrome patients with 5 to 6 mm of instability should be restricted from participating in sports that carry a risk of stress to the head and neck. In 1995, the AAP retired this recommendation and issued the following statement (109): “From the available scientific evidence, it is reasonable to conclude that lateral plain radiographs of the cervical spine are of potential but unproven value in detecting patients at risk for developing spinal cord injury during sports participation.” Current opinion is that in asymptomatic children, yearly examinations to detect any neurologic symptoms or signs of myelopathy are more predictive of progressive myelopathy or neurologic injury than are screening x-rays (182). Evaluation of lateral cervical spine x-rays in full flexion and full extension is still required before participation in sports considered by the Special Olympics to have potential risk: certain activities that axially load the head in flexion, such as gymnastics, diving, and soccer (187). Davidson (182) found that neurologic signs were more predictive of impending dislocation than the x-ray criteria. Studies have shown that by adolescence, the frequency of atlantoaxial instability approaches 10% to 30% (109,174,184,188,189). It also appears that 12% (174) to 16% (183) of children with Down syndrome who have instability develop neurologic signs and symptoms.
Surgical stabilization is indicated for patients with translation of more than 10 mm. In patients with less than 10 mm of translation and a neurologic deficit or history of neurologic symptoms, surgical stabilization also may be indicated. Once surgical stabilization is needed, the treating physician must understand the increased risk of complications (i.e., pseudarthrosis) in this patient population. Segal et al (188) reported a high complication rate after posterior arthrodesis of the cervical spine in patients with Down syndrome. Six of 10 patients developed resorption of the bone graft and associated pseudarthrosis. Other complications in this patient population after attempted posterior arthrodesis were wound infection, dehiscence of the operative site, instability of adjacent motion segments, and neurologic sequelae (180).
Atlantoaxial Rotatory Subluxation
Atlantoaxial rotatory subluxation is a common cause of childhood torticollis. This condition is known by several names, such as rotatory dislocation, rotatory displacement, rotatory subluxation, and rotatory fixation. Atlantoaxial rotatory subluxation probably is the most accepted term used, except for long-standing cases (3 months), which are called rotatory fixation.
A significant amount of motion occurs at the atlantoaxial joint; half of the rotation of the cervical spine occurs there. Through this range of motion at the C1-C2 articulation, some children develop atlantoaxial rotatory subluxation. The two most common causes are trauma and infection; the most common cause is an upper respiratory infection (Grisel syndrome) (190). Subluxation also can occur after a retropharyngeal abscess, tonsillectomy, pharyngoplasty, or trivial trauma. There is free blood flow between the veins and lymphatics draining the pharynx and the periodontoid plexus (75). Any inflammation of these structures can lead to attenuation of the synovial capsule or transverse ligament or both, with resulting instability. Another potential etiologic factor is the shape of the superior facets of the axis in children. Kawabe et al (191) showed that the facets are smaller and more steeply inclined in children than in adults. A meniscus-like synovial fold was found between C1 and C2 that could prohibit reduction after displacement has occurred.
Fielding and Hawkins (155) classified atlantoaxial rotatory displacements into four types based on the direction and degree of rotation and translation (Fig. 18-36). Type 1 is a unilateral facet subluxation with an intact transverse ligament. This is the most common and benign type. Type 2 is a unilateral facet subluxation with anterior displacement of 3 to 5 mm. The unilateral anterior displacement of one of the lateral masses may indicate an incompetent transverse ligament with potential instability. Type 3 is bilateral anterior facet displacement with more than 5 mm of anterior displacement. This type is associated with deficiencies of the transverse and secondary ligaments, which can result in significant narrowing of the space available

for the cord at the atlantoaxial level. Type 4 is an unusual type in which the atlas is displaced posteriorly. This usually is associated with a deficient dens. Although types 3 and 4 are rare, neurologic involvement may be present or instantaneous death can occur. Both types must be managed with great care.
FIGURE 18-36 Classification of rotary displacement. (Reprinted from Fielding JW, Hawkins RJ. Atlantoaxial rotary fixation. J Bone Joint Surg [Am] 1977;59:37; with permission.)
Signs and Symptoms
Clinical findings include neck pain, headache, and a cock-robin position of rotating to one side as well as lateral flexion to the other (Fig. 18-37). When rotatory subluxation is acute, the child resists attempts to move the head and has pain with any attempts at correction. Usually the child is able to make the deformity worse but cannot correct it. Associated muscle spasms of the sternocleidomastoid muscle occur predominantly on the side of the long sternocleidomastoid muscle in an attempt to correct the deformity. If the deformity becomes fixed, the pain subsides but the torticollis and the decreased range of motion will persist (155). If rotatory fixation has been present for a long time in a small child, plagiocephaly is sometimes noted. Neurologic abnormalities are extremely rare, although a few cases have been reported.
X-Ray Findings
Adequate x-rays may be difficult to obtain because of the associated torticollis and difficulty in positioning the head and neck. Anteroposterior and open-mouth odontoid views should be taken with the shoulders flat and the head in as neutral a position as possible (192). Lateral masses that have rotated forward appear wider and closer to the midline, whereas the opposite lateral mass appears narrower and farther away from the midline on this view. One of the facet joints may be obscured because of apparent overlapping. The distance between the lateral mass and the dens also will be asymmetric. On the lateral view, the lateral facet appears anterior and usually appears wedge-shaped instead of the normal oval shape. The posterior arches of the atlas may fail to superimpose because of head tilt, giving the appearance of fusion of C1 to the occiput (occipitalization). Flexion and extension lateral views are recommended to exclude instability.
FIGURE 18-37 Child with rotary subluxation of C1 on C2. Note the direction of head tilt and rotation of the neck.
Cineradiography has been used for the evaluation of atlantoaxial rotatory subluxation (153,157,193). This technique is limited in the acute stage because pain restricts the motion necessary for a satisfactory study. With atlantoaxial rotatory fixation, cineradiography may be helpful in confirming the diagnosis by showing that the atlas and axis are rotating as a unit. However, this technique requires high radiation exposure and generally has been replaced by CT scanning (140,144,157,194,195). CT should be performed with the head and body positioned as close to neutral as possible. This will show a superimposition of C1 on C2 in a rotated position and will allow the degree and amount of malrotation to be quantified. Some researchers have recommended dynamic CT scans taken with

the patient looking to the right and the left to diagnose rotatory fixation (196). Three-dimensional CT scans also are helpful in identifying rotatory subluxation (146). MRI demonstrates more soft tissue detail, such as associated spinal cord compression and underlying vertebral or soft tissue infections (Fig. 18-38) (197).
Differential Diagnoses
Differential diagnoses include torticollis caused by ophthalmologic problems, sternocleidomastoid tightness from muscular torticollis, brain stem or posterior fossa tumors or abnormalities, congenital vertebral anomalies, and infections of the vertebral column.
FIGURE 18-38 A, B. Odontoid view and lateral cervical spine x-ray of rotary subluxation of C1 on C2. C. Note the asymmetry on the open-mouth odontoid view. D. CT and CT reconstruction documenting rotary subluxation.
Treatment depends on the duration of the symptoms (196). Many patients probably never receive medical treatment, because symptoms may be mild and the subluxation may reduce spontaneously over a few days before medical attention is sought. If rotatory subluxation has been present for a week or less, a soft collar, anti-inflammatory medication, and an exercise program are indicated. If this fails to produce improvement and the symptoms persist for more than a week, head halter traction should be initiated. This can be done either at home or in the hospital, depending on the social situation and the severity of symptoms. Muscle relaxants and analgesics also may be needed. Phillips and Hensinger (196) found that if rotatory subluxation

was present for less than 1 month, head halter traction and bed rest were usually sufficient to relieve symptoms. If the subluxation has been present for longer than a month, successful reduction is not very likely (116). However, halo traction can still be used to try to reduce the subluxation. The halo allows increased traction weight to be applied without interfering with opening of the jaw or causing skin pressure on the mandible. While the traction is being applied, active rotation to the right and left should be encouraged. Once the atlantoaxial rotatory subluxation has been reduced, motion has been restored, and the reduction is documented by CT scan, the patient is maintained in a halo vest for 6 weeks. If reduction cannot be maintained, posterior atlantoaxial arthrodesis is recommended. Even though internal rotation and alignment of the atlas and axis may not be restored, successful fusion should result in the appearance of normal head alignment by relieving the muscle spasms that occurred in response to the malrotation. Posterior arthrodesis also is recommended if any signs of instability or neurologic deficits secondary to the subluxation are present, if the deformity has been present for more than 3 months, or if conservative treatment of 6 weeks of immobilization has failed.
Hangman’s Fracture
Bilateral spondylolisthesis of C2, or hangman’s fractures, also may occur in children (197). The mechanism of injury is forced hyperextension. Most reports of this injury have been in children under the age of 2 years (195,198,199,200,201,202,203,204). This injury probably occurs more frequently in this age group because of the disproportionately large head, poor muscle control, and hypermobility. The possibility of child abuse also must be considered. Patients present with neck pain and resist any movement of the head and neck. There should be a positive history of trauma (Fig. 18-39).
X-rays reveal a lucency anterior to the pedicles of the axis, usually with some forward subluxation of C2 on C3. One must be sure this is a fracture and not a persistent synchondrosis of the axis. Several authors (205,206,207,208) have reported similar cases of persistent synchondrosis of the axis. CT scans showed the defect to be at the level of the neurocentral chondrosis. Later films showed ossification within the synchondrosis gap.
FIGURE 18-39 Lateral x-ray of patient with traumatic C2 spondylolisthesis (hangman’s fracture).
Treatment should be symptomatic with immobilization in a Minerva cast, halo, or cervical orthosis for 8 to 12 weeks. Traction is not needed to reduce this fracture and may even produce potentially dangerous distraction. Pizzutillo et al (195) reported that four of five patients healed with immobilization. If union does not occur, posterior arthrodesis or anterior arthrodesis can be performed to stabilize this fracture.
Fractures and dislocations involving C3 through C7 are rare in children and infants (90,209,210,211) and usually occur in teenagers or older children. Lower cervical spine injuries in children as opposed to those in adults can occur through the cartilaginous endplate (130). The endplate may break completely through the cartilaginous portion (Salter type I) or may exit through the bony edge (Salter type II). Usually the inferior endplate fractures because of the protective effect of the uncinate processes of the superior endplate (13).
Posterior Ligamentous Disruption
Posterior ligamentous disruption can occur with a flexion or distraction injury to the cervical spine. The patient usually has point tenderness at the injury site and complains of neck pain. Initial x-rays may be normal except for loss of normal cervical lordosis. This may be a normal finding in young children but should be evaluated for possible ligamentous injury in an adolescent. Widening of the posterior interspinous distance is suggestive of this injury. MRI may be helpful in documenting ligamentous damage.
With posterior ligamentous disruption, gradual displacement

of one segment on the other can occur, and secondary adaptive changes in the growing spine may make reduction difficult. Posterior ligamentous injuries should be protected with an extension orthosis, and patients should be followed closely for the development of instability. If signs of instability are present, then a posterior arthrodesis should be performed.
Compression Fractures
Compression fractures, the most common fractures of the subaxial spine in children, are caused by flexion and axial loading that results in loss of vertebral body height. This can be detected on a lateral x-ray. Because the vertebral disks in children are more resilient than the vertebral bodies, the bone is more likely to be injured. Compression fractures are stable injuries and heal in children in 3 to 6 weeks. Many compression fractures may be overlooked because of the normal wedge shape of the vertebral bodies in young children. Immobilization in a cervical collar is recommended for 3 to 6 weeks. Flexion and extension films to confirm stability should be obtained 2 to 4 weeks after injury. In children under 8 years of age, the vertebral body may reconstitute itself with growth, although Schwarz et al (212) reported that kyphosis of more than 20 degrees may not correct with growth. Associated injuries can include anterior teardrop, laminar, and spinous process fractures.
Unilateral and Bilateral Facet Dislocations
Unilateral facet dislocations and bilateral facet dislocations are the second most common injuries in the subaxial spine in children. Most occur in adolescents and are similar to adult injuries. The diagnosis usually can be made on anteroposterior and lateral x-rays. In children the so-called perched facet is a true dislocation. The cartilaginous components are overlapped and locked. On the x-ray, the facet appears perched because the overlapped cartilage cannot be seen. Unilateral facet dislocation is treated with traction and reduction. If reduction cannot be easily obtained, open reduction and arthrodesis are indicated. Complete bilateral facet dislocation, although rare, is more unstable and has a higher incidence of neurologic deficit (Fig. 18-40). Treatment consists of reduction and stabilization with a posterior arthrodesis.
Burst Fractures
Although rare, burst fractures can occur in children. These injuries are caused by an axial load. X-ray evaluation should consist of anteroposterior and lateral views. CT scans aid in detecting any spinal canal compromise from retropulsed fracture fragments and occult laminar fractures. If no neurologic deficit or significant canal compromise is present, then treatment consists of traction followed by halo immobilization. Anterior arthrodesis rarely is recommended in pediatric patients, except in a patient with a burst fracture and significant canal compromise (213). Anterior arthrodesis destroys the anterior growth potential; as posterior growth continues, a kyphotic deformity may occur (Fig. 18-41). In older children and adolescents, anterior instrumentation can be used for stabilization.
Spondylolysis and Spondylolisthesis
Spondylolysis and spondylolisthesis of C2 through C6 have been reported. These injuries can occur from either a hyper-extension or flexion axial loading injury. Associated anterosuperior avulsion or compression fractures of the vertebral body may occur. The diagnosis usually is made on plain x-rays that show a fracture line through the pedicles. Oblique views may be necessary to better identify the fracture line. CT scanning may be useful in differentiating an acute fracture from a normal synchondrosis. Treatment consists of immobilization in a cervical orthosis or halo brace. Surgical stabilization is recommended only for truly unstable fractures or nonunions. Neurologic involvement is rare.
Operative Treatment
Posterior Arthrodesis
General anesthesia is administered with the patient supine (Fig. 18-42). The patient is turned prone on the operating table, with care taken to maintain traction and proper alignment of the head and neck. The head may be positioned in a head rest or maintained in skeletal traction. X-rays are obtained to confirm adequate alignment of the vertebrae and to localize the vertebrae to be exposed. Extension of the fusion mass can occur when extra vertebrae or spinous processes are exposed in the cervical spine. A midline incision is made over the chosen spinous processes, and the spinous process and lamina are exposed subperiosteally to the facet joints. If the spinous process is large enough, a hole is made in the base of the spinous process with a towel clip or Lewin clamp. An 18-gauge wire is passed through this hole, looped over the spinous process, and passed through the hole again. A similar hole is made in the base of the spinous process of the inferior vertebra to be fused, and the wire is passed through this vertebra. The wire is then passed through this hole, looped under the inferior aspect of the spinous process, and then passed back through the same hole. The wire is tightened and corticocancellous bone grafts are placed along the exposed lamina and spinous processes. The wound is closed in layers. If the spinous process is too small to pass wires, then an in situ arthrodesis can be performed and external immobilization used.
Hall et al (214) used a 16-gauge wire and threaded Kirschner wires. The threaded Kirschner wires are passed through the bases of the spinous processes of the vertebrae to be fused. This is followed by a figure-of-eight wiring with a 16-gauge wire (Fig. 18-43). After tightening the wire about the Kirschner wires, strips of corticocancellous and cancellous bone are packed over the posterior arches of the vertebrae to be fused.
In older children and adolescents, lateral mass plates or screw-and-rod systems can be used in the lower cervical spine. The instrumentation should be of appropriate size to match the size of the child’s cervical spine.
FIGURE 18-40 A, B. Lateral x-ray of a patient with so-called perched facets, demonstrating a facet dislocation. C, D. Lateral and anteroposterior x-rays after reduction and posterior arthrodesis.
FIGURE 18-41 Anteroposterior and lateral x-rays and CT scan of patient with a minimally displaced burst fracture of C5.
FIGURE 18-42 Technique of posterior arthrodesis in subaxial spine levels C3-C7. A. A hole is made in the spinous process of the vertebrae to be fused. B. An 18-gauge wire is passed through both holes and around the spinous processes. C. The wire is tightened. D. Corticocancellous bone grafts are placed. (Redrawn from Murphy MJ, Southwick WO. Posterior approaches and fusions. In: Cervical Spine Research Society. The cervical spine. Philadelphia: JB Lippincott, 1983; with permission.)
FIGURE 18-43 Alternative fixation method for posterior arthrodesis of C3-C7. A 16-gauge wire is placed in a figure-of-eight pattern around two threaded Kirschner wires passed through the bases of the spinous processes of the vertebrae to be fused. (Reprinted from Hall JE, Simmons ED, Danylchuk K, et al. Instability of the cervical spine and neurological involvement in Klippel-Feil syndrome: a case report. J Bone Joint Surg [Am] 1990;72:460; with permission.)



Posterior Arthrodesis with Lateral Mass Screw Fixation
Several techniques of lateral mass screw fixation for the lower cervical spine have been described. They differ primarily in the entry points for the screws and in the trajectory of screw placement, which yield different exit points (215,216).
Roy-Camille Technique
The entry point for the screw is at the center of the rectangular posterior face of the lateral mass or can be measured 5 mm medial to the lateral edge and midway between the facet joints (Fig. 18-44A). The drill is directed perpendicular to the posterior wall of the vertebral body with a 10-degree lateral angle (see Fig. 18-44B). This trajectory establishes an exit point slightly lateral to the vertebral artery and below the exiting nerve root. The lateral mass depth from C3 to C6 ranges from 6 to 14 mm in men (average 8.7 mm) and 6 to 11 mm in women (average 7.9 mm). An adjustable drill guide set to a depth of 10 to 12 mm is used to prevent penetration beyond the anterior cortex. The depth can be gradually and safely increased if local anatomy permits. If the additional 20% of pullout strength with bicortical fixation is desired, the exit point should be at the junction of the lateral mass and the transverse process. Lateral fluoroscopic imaging makes it easier to choose the optimal trajectory and avoid penetration of the subjacent facet joint (see Fig. 18-44C),

which is especially important at the caudal level of fixation because this joint should be included in the fusion.
FIGURE 18-44 Roy-Camille technique of lateral mass screw insertion. A. Entry point for screw insertion. B. Drill is directed perpendicular to posterior wall of vertebral body with a 10-degree lateral angle. C. Final screw position. (Reprinted from Heller JG, Jeffords P. Internal fixation of the cervical spine. Posterior instrumentation of the lower cervical spine. In: Frymoyer JW, Wiesel SW, eds. The adult and pediatric spine, 3rd ed. Philadelphia: Lippincott Williams & Wilkins, 2004; with permission.)
Magerl Technique
The entry point for the screw is 1 mm medial and rostral (proximal) to the center point of the posterior surface of the lateral mass (Fig. 18-45A). It is oriented at a 45- to 60-degree rostral angle, parallel to the adjacent facet joint articular surface, and at a 25-degree lateral angle (see Fig. 18-45B). This trajectory establishes an exit point lateral to the vertebral artery and above the exiting nerve root while engaging the lateral portion of the ventral cortex of the superior articular facet (see Fig. 18-45C). The proper trajectory for this technique is more difficult to achieve that in the Roy-Camille technique. The prominence of the thorax can impede proper alignment of

the drill and guide, risking injury to the nerve root if the second cortex is penetrated. The depth of penetration at this angle is approximately 18 mm, compared to 14 mm with the Roy-Camille technique, which has some implications for purchase strength and mode of screw failure.
FIGURE 18-45 Magerl technique of lateral mass screw insertion. A. Entry point for screw insertion. B. Drill is directed at a 25-degree lateral angle. C. Final screw position. (Reprinted from Heller JG, Jeffords P. Internal fixation of the cervical spine. Posterior instrumentation of the lower cervical spine. In: Frymoyer JW, Wiesel SW, eds. The adult and pediatric spine, 3rd ed. Philadelphia: Lippincott Williams & Wilkins, 2004; with permission.)
1. Anderson PA, Montesano PX. Morphology and treatment of occipital condyle fractures. Spine 1988;S13:731–736.
2. Anderson JM, Schutt AH. Spinal injury in children: a review of 156 cases seen from 1950 through 1978. Mayo Clin Proc 1980;S55:499–504.
3. Hadley MN, Zabramski JM, Browner CM, et al. Pediatric spinal trauma: review of 122 cases of spinal cord vertebral column injuries. J Neurosurg 1988;S68:18–24.
4. Haffner DL, Hoffer MM, Wiedebusch R. Etiology of children’s spinal injuries at Rancho Los Amigos. Spine 1993;S18:679–684.
5. Herzenberg JE, Hensinger RN. Pediatric cervical spine injuries. Trauma Q 1989;S5:73–81.
6. Hubbard DD. Injuries of the spine in children and adolescents. Clin Orthop 1974;S100:56–65.
7. Kewalramani LS, Kraus JF, Sterling HM. Acute spinal-cord lesions in a pediatric population: epidemiological and clinical features. Paraplegia 1980;S18:206–219.
8. Lebwohl NH, Eismont FJ. Cervical spine injuries in children. In: Weinstein SL, ed. The pediatric spine: principles and practice. Philadelphia: Lippincott Williams & Wilkins, 2001.
9. Pang D, Wilberger JE. Spinal cord injury without radiologic abnormalities in children. J Neurosurg 1982;S57:114–129.
10. Ruge JR, Sinson GP, McLone DG, et al. Pediatric spinal injury: the very young. J Neurosurg 1988;S68:25–30.
11. Stauffer ES, Mazur JM. Cervical spine injuries in children. Pediatr Ann 1982;S11:502–511.
12. McGrory BJ, Klassen RA, Chao EY, et al. Acute fracture and dislocations of the cervical spine in children and adolescents. J Bone Joint Surg [Am] 1993;S75:988–995.
13. Aufdermaur M. Spinal injuries in juveniles: necropsy findings in twelve cases. J Bone Joint Surg [Br] 1974;S56:513–519.
14. Pang D, Pollack IF. Spinal cord injury without radiologic abnormality in children: the SCIWORA syndrome. J Trauma 1989;S29:654–664.
15. Ghatan S, Ellenbogen RG. Pediatric spine and spinal cord injury after inflicted trauma. Neurosurg Clin North Am 2002;13:227–233.
16. Brooks AL, Jenkins EB. Atlantoaxial arthrodesis by the wedge compression method. J Bone Joint Surg [Am] 1978;S60:279–290.
17. Zuckerbraun BS, Morrison K, Baines B, et al. Effect of age on cervical spine injuries in children after motor vehicle collisions: effectiveness of restraint devices. J Pediatr Surg 2004;S39:483–486.
18. Torg B, Das M. Trampoline and minitrampoline injuries to the cervical spine. Clin Sports Med 1985;S4:45–60.
19. Birney TJ, Hanley EN. Traumatic cervical spine injuries in childhood and adolescence. Spine 1989;S14:1277–1282.
20. Evans DL, Bethem D. Cervical spine injuries in children. J Pediatr Orthop 1989;S9:563–568.
21. Eleraky MA, Theodore N, Adams M, et al. Pediatric cervical spine injuries: report of 102 cases and review of the literature. J Neurosurg 2000;S92(1 Suppl):12–17.
22. Kilfoyle RM, Foley JJ, Norton PL. Spine and pelvic deformity in childhood and adolescent paraplegia. J Bone Joint Surg [Am] 1965;S47:659–682.
23. Nitecki S, Moir CR. Predictive factors of the outcome of traumatic cervical spine fracture in children. J Pediatr Surg 1994;S29:1409–1411.
24. Patel JC, Tepas JJ III, Mollitt DL, Pieper P. Pediatric cervical spine injuries: defining the disease. J Pediatr Surg 2001;S36:373–376.
25. Orenstein JB, Klein BL, Gotschall CS, et al. Age and outcome in pediatric cervical spine injury: 11-year experience. Pediatr Emerg Care 1994;S10:132–137.
26. Orenstein JB, Klein BL, Oschensclager DW. Delayed diagnosis of pediatric cervical spine injury. Pediatrics 1992;S89:1185–1188.
27. Hamilton MG, Myles ST. Pediatric spinal injury. Review of 61 deaths. J Neurosurg 1988;S77:705–708.
28. Caffey J. The whiplash shaken infant syndrome. Pediatrics 1974;S54:396–403.
29. Cattell HS, Filtzer DL. Pseudosubluxation and other normal variations in the cervical spine in children. J Bone Joint Surg [Am] 1965;S47:1295–1309.
30. Lawson JP, Ogden JA, Bucholz RW, et al. Physeal injuries of the cervical spine. J Pediatr Orthop 1987;S7:428–435.
31. Rachesky I, Boyce WT, Duncan B, et al. Clinical prediction of cervical spine injuries in children: radiographic abnormalities. Am J Dis Child 1987;S141:199–201.
32. Bohn D, Armstrong D, Becker L, et al. Cervical spine injuries in children. J Trauma 1990;S30:463–469.
33. Committee on Pediatric Orthopaedics. Trauma of the Cervical Spine. Position Statement, American Academy of Orthopaedic Surgeons, 1990.
34. Pennecot GF, Gouraud D, Hardy JR, et al. Roentgenographical study of the stability of the cervical spine in children. J Pediatr Orthop 1984;S4:346–352.
35. Hensinger RN, Lang JE, MacEwen GD. Klippel-Feil syndrome: a constellation of associated anomalies. J Bone Joint Surg [Am] 1974;S56:1246–1252.
36. Klippel M, Feil A. Anomalies de la collone vertebrale par absence des vertebres cervicales; avec cage thoraque remontant jusqu’ala bas du crane. Bull Soc Anat Paris 1912;S87:185.
37. Hensinger RN, Fielding JW, Hawkins RJ. Congenital anomalies of the odontoid process. Orthop Clin North Am 1978;S9:901–912.
38. Kokoska ER, Keller MS, Rallo MC, Weber TR. Characteristics of pediatric cervical spine injuries. J Pediatr Surg 2001;S36:100–105.
39. Huerta C, Griffith R, Joyce SM. Cervical spine stabilization in pediatric patients. Evaluation of current techniques. Ann Emerg Med 1987;S16:1121–1126.
40. Schwartz GR, Wright SW, Fein JA, et al. Pediatric cervical spine injury sustained in falls from low heights. Ann Emerg Med 1997;S30:249–252.
41. Fuchs S, Barthel MJ, Flannery AM, et al. Cervical spine fractures sustained by young children in forward-facing car seats. Pediatrics 1989;S84:348–354.
42. Hoy GA, Cole WG. The paediatric cervical seat belt syndrome. Injury 1993;S24:297–299.
43. Viccellio P, Simon H, Pressman BD, et al. A prospective multicenter study of cervical spine injury in children. Pediatrics 2001;S108:E20.
44. Anderson LD, Smith BL, DeToree J, et al. The role of polytomography in the diagnosis and treatment of cervical spine injuries. Clin Orthop Rel Res 1982;S165:64–68.
45. Dietrich AM, Ginn-Pease ME, Bartkowski HM, et al. Pediatric cervical spine fractures: predominately subtle presentation. J Pediatr Surg 1991;S26:995–1000.
46. Lally KP, Senak M, Hardin WD, et al. Utility of the cervical spine radiograph in pediatric trauma. Am J Surg 1989;S158:540–542.
47. Letts M, Kaylor D, Gouw G. A biomechanical study of halo fixation in children. J Bone Joint Surg [Br] 1987;S70:277–279.
48. Bachulis BL, Long WB, Hynes GD, et al. Clinical indications for cervical spine radiographs in the traumatized patient. Am J Surg 1987;S153:473–477.
49. Shaffer MA, Doris PE. Limitation of the cross-table lateral view in detecting cervical spine injuries: a retrospective review. Ann Emerg Med 1981;S10:508–513.
50. Ralston ME, Chung K, Barnes PD, et al. Role of flexion-extension radiographs in blunt pediatric cervical spine injury. Acad Emerg Med 2001;S8:237–245.
51. Donahue D, Muhlbauer M, Kaufman R, et al. Childhood survival of atlanto-occipital dislocation: underdiagnosis, recognition, treatment, and review of the literature. Pediatr Neurosurg 1994;S21:105–111.
52. Bulas DI, Fitz CR, Johnson DL. Traumatic atlanto-occipital dislocation in children. Radiology 1993;S188:155–158.
53. de Beer JD, Hoffman EB, Kieck CF. Traumatic atlantoaxial subluxation in children. J Pediatr Orthop 1990;S10:397–400.
54. Swischuk EH Jr, Rowe ML. The upper cervical spine in health and disease. Pediatrics 1952;S10:567–572.
55. Allington JJ, Zembo M, Nadell J, Bowen JR. C1-C2 posterior soft tissue injuries with neurologic impairment in children. J Pediatr Orthop 1990;S10:596–601.
56. Walsh JW, Stevens DB, Young AB. Traumatic paraplegia in children without contiguous spinal fracture or dislocation. Neurosurgery 1983;S12:439–445.
57. Kuhns LR, Strouse PJ. Cervical spine standards for flexion radiograph interspinous distance ratios in children. Acta Radiol 2000;S7:615–619.
58. Hedequist DJ, Emans JB. The correlation of preoperative three-dimensional computed tomography reconstructions with operative findings in congenital scoliosis. Spine 2003;S28:2531–2534.
59. Keenan HT, Hollingshead MC, Chung CJ, et al. Using CT of the cervical spine for early evaluation of pediatric patients with head trauma. AJR Am J Roentgenol 2001;S177:1405–1409.
60. Adlegais KM, Grossman DC, Langer SG, et al. Use of helical computed tomography for imaging the pediatric cervical spine. Acad Emerg Med 2004;S11:228–236.
61. Flynn JM, Closkey RF, Mahboubi S, et al. Role of magnetic resonance imaging in the assessment of pediatric cervical spine injuries. J Pediatr Orthop 2002;S22:573–577.
62. Parbhoo AH, Govender S, Corr P. Vertebral artery injury in cervical spine trauma. Injury 2001;S32:565–568.
63. Herzenberg JE, Hensinger RN, Dedrick DK, et al. Emergency transport and positioning of young children who have an injury of the cervical spine: the standard backboard may be hazardous. J Bone Joint Surg [Am] 1989;S71:15–22.
64. Nypaver M, Treloar D. Neutral cervical spine positioning in children. Ann Emerg Med 1994;S23:208–211.
65. Curran C, Dietrich AM, Bowman MJ, et al. Pediatric cervical-spine immobilization: achieving neutral position? J Trauma 1995;S39:729–732.
66. Millington PJ, Ellingsen JM, Hauswirth BE, et al. Thermoplastic Minerva body jacket=ma practical alternative to current methods of cervical spine stabilization. Phys Ther 1987;S67:223–225.
67. Dormans JP, Criscitiello AA, Drummond DS, et al. Complications in children managed with immobilization in a halo vest. J Bone Joint Surg [Am] 1995;S77:1370–1373.
68. Baum JA, Hanley EN Jr, Pullekines J. Comparison of halo complications in adults and children. Spine 1989;S14:251–252.
69. Mubarak SJ, Camp JF, Vueltich W, et al. Halo application in the infant. J Pediatr Orthop 1989;S9:612–614.
70. Garfin SR, Roux R, Botte MJ, et al. Skull osteology as it affects halo pin placement in children. J Pediatr Orthop 1986;S6:434–436.
71. Copley LA, Dormans JP, Pepe MD, et al. Accuracy and reliability of torque wrenches used for halo application in children. J Bone Joint Surg [Am] 2003;S85:2199–2204.
72. Copley LA, Pepe MD, Tan V, et al. A comparison of various angles of halo pin insertion in an immature skull model. Spine 1999;S24:1777–17800.
73. Burke DC. Spinal cord trauma in children. Paraplegia 1971;S9:1–14.
74. Taylor AR. The mechanism of injury to the spinal cord in the neck without damage to the vertebral column. J Bone Joint Surg [Br] 1951;S33:453–547.

75. Parke WW, Rothman RH, Brown MD. The pharyngovertebral veins: an anatomical rationale for Grisel’s syndrome. J Bone Joint Surg [Am] 1984;S66:568–574.
76. Yngve DA, Harris WP, Herndon WA, et al. Spinal cord injury without osseous spine fracture. J Pediatr Orthop 1988;S8:153–159.
77. Menezes AH, Ryken JC. Craniovertebral junction abnormalities. In: Weinstein SL, ed. The pediatric spine: principles and practice. New York: Raven, 1994.
78. Annis JA, Finlay DB, Allen MJ, et al. A review of cervical-spine radiographs in casualty patients. Br J Radiol 1987;S60:1059–1061.
79. Hanson JA, Deliganis AV, Baxter AB, et al. Radiologic and clinical spectrum of occipital condyle fractures: retrospective review of 107 consecutive fractures in 95 patients. AJR Am J Roentgenol 2002;S178:1261–1268.
80. Hensinger RN, DeVito PD, Ragsdale CG. Changes in the cervical spine in juvenile rheumatoid arthritis. J Bone Joint Surg [Am] 1986;S68:189–198.
81. Ranjith RK, Mullett JH, Burke TE. Hangman’s fracture cause by suspected child abuse. A case report. J Pediatr Orthop B 2002;S11:329–332.
82. Bracken MB. Treatment of acute spinal cord injury with methylprednisolone: results of a multi-center randomized clinical trial. J Neurotrauma 1991;S8(suppl 1):47–50.
83. Bracken MB. Pharmacological treatment of acute spinal cord injury: current status and future projects. J Emerg Med 1993;S11(suppl 1):43–48.
84. Bracken MB, Shepard MJ, Collins WF, et al. A randomized controlled trial of methylprednisolone or naloxone in the treatment of acute spinal cord injury: results of the Second National Spinal Cord Injury Study. N Engl J Med 1990;S322:1405–1411.
85. Bracken MB, Shepard MJ, Collins WF, et al. Methylprednisolone and naloxone treatment after acute spinal cord injury: one-year follow-up data. Results of the Second National Acute Spinal Cord Injury Study. J Neurosurg 1992;S76:23–31.
86. Geisler FH, Dorsey FC, Coleman WP. Recovery of motor function after spinal cord injury=ma randomized, placebo-controlled trial with GM-1 ganglioside [erratum]. N Engl J Med 1991;S325:1669–1670.
87. Geisler FH, Dorsey FC, Coleman WP. Recovery of motor function after spinal cord injury=ma randomized, placebo-controlled trial with GM-1 ganglioside. N Engl J Med 1991;S324:1829–1838.
88. Geisler FH, Dorsey FC, Coleman WP. GM-1 ganglioside in human spinal cord injury. J Neurotrauma 1992;S9(suppl 1):407–416.
89. Geisler FH, Dorsey FC, Coleman WP. Past and current clinical studies with GM-1 ganglioside in acute spinal cord injury. Rev Ann Emerg Med 1993;S22:1041–1047.
90. Sherk HH, Schut L, Lane J. Fractures and dislocations of the cervical spine in children. Orthop Clin North Am 1976;S7:593–604.
91. Yasuoko F, Peterson H, MacCarty C. Incidence of spinal column deformity after multiple level laminectomy in children and adults. J Neurosurg 1982;S57:441–445.
92. Mayfield JK, Erkkila JC, Winter RB. Spine deformities subsequent to acquired childhood spinal cord injury. Orthop Trans 1979;S3:281–282.
93. Sim F, Svien H, Bickel W, et al. Swan neck deformity following extensive cervical laminectomy. J Bone Joint Surg [Am] 1974;S56:564–580.
94. Banniza von Bazan UK, Paeslack V. Scoliotic growth in children with acquired paraplegia. Paraplegia 1977;S15:65–73.
95. Bedbrook GM. Correction of scoliosis due to paraplegia sustained in pediatric age group. Paraplegia 1977;S15:90–96.
96. Burke DC. Traumatic spinal paralysis in children. Paraplegia 1971;S9:268–276.
97. Farley FA, Hensinger RN, Herzenberg JE. Cervical spinal cord injury in children. J Spinal Disord 1992;S5:410–416.
98. Norman MG, Wedderburn LC. Fetal spinal cord injury with cephalic delivery. Obstet Gynecol 1973;S42:355–358.
99. Tawbin A. CNS damage in the human fetus and newborn infant. Am J Dis Child 1951;S33:543–547.
100. Maekawa K, Masaki T, Kokubun Y. Fetal spinal cord injury secondary to hyperextension of the neck: no effect of caesarean section. Dev Med Child Neurol 1976;S18:228–232.
101. Bresnan MJ, Abroms IF. Neonatal spinal cord transection secondary to intrauterine hyperextension of neck in breech presentation. J Pediatr 1974;S84:734–737.
102. Conry BG, Hall CM. Cervical spine fractures and rear car seat restraints. Arch Dis Child 1987;S62:1267–1268.
103. Givens T, Polley KA, Smith GF, Hardin WD Jr. Pediatric cervical spine injury: a three-year experience. J Trauma 1996;S41:310–314.
104. Shulman ST, Madden JD, Esterly JR, et al. Transection of the spinal cord. A rare obstetrical complication of cephalic delivery. Arch Dis Child 1971;S46:291–294.
105. Swischuk LE. Spine and spinal cord trauma in the battered child syndrome. Radiology 1969;S92:733–738.
106. Momjian S, Dehdashti AR, Kehrli P, et al. Occipital condyle fractures in children: case report and review of the literature. Pediatr Neurosurg 2003;S38:265–270.
107. Noble ER, Smoker WRK. The forgotten condyle: the appearance, morphology, and classification of occipital condyle fractures. AJNR Am J Neuroradiol 1996;S17:507–513.
108. Tuli S, Tator CH, Fehlings MG, Mackay M. Occipital condyle fractures. Neurosurgery 1997;S41:368–377.
109. American Academy of Pediatrics Committee on Sports Medicine and Fitness. Atlanto-axial instability in Down syndrome: subject review. Pediatrics 1995;S96(1 pt 1):151–154.
110. Bohlman HH. Acute fractures and dislocations of the cervical spine. J Bone Joint Surg [Am] 1969;S61:1119–1142.
111. Bucholz RW, Burkhead WZ. The pathological anatomy of fatal atlanto-occipital dislocations. J Bone Joint Surg [Am] 1979;S61:248–250.
112. Evarts CM. Traumatic occipito-atlanto dislocation. J Bone Joint Surg [Am] 1970;S52:1653–1660.
113. Papadopoulos SM, Dickman CA, Sonntag VKH, et al. Traumatic atlantooccipital dislocation with survival. Neurosurgery 1991;S28:574–579.
114. Sponseller PD, Herzenberg JE. Cervical spine injuries in children. In: The cervical spine. Philadelphia: Lippincott-Raven, 1998:357–371.
115. Bernini EP, Elefante R, Smaltino F, Tedeschi G. Angiographic study on the vertebral artery in cases of deformities of the occipitocervical joint. AJR Am J Roentgenol 1969;S107:526–529.
116. Burkus JK, Deponte RJ. Chronic atlantoaxial rotatory fixation: correction by cervical traction, manipulation, and branching. J Pediatr Orthop 1986;S6:631–635.
117. Collato PM, Demuth WW, Schwentker EP, et al. Traumatic atlanto-occipital dislocation. J Bone Joint Surg [Am] 1986;S68:1106–1109.
118. Harmanli O, Kaufman Y. Traumatic atlanto-occipital dislocation with survival. Surg Neurol 1993;S39:324–330.
119. Hosono N, Yonenbou K, Kazuyoshi K, et al. Traumatic anterior atlanto-occipital dislocation. Spine 1993;S18:786–790.
120. Powers B, Milber MD, Kramer RS, et al. Traumatic anterior occipital dislocation. Neurosurgery 1979;S4:12–17.
121. Bundschumb CV, Allen JB, Ross M, et al. Magnetic resonance imaging of suspected atlanto-occipital dislocation. Spine 1992;S17:245–248.
122. Farley FA, Graziano GP, Hensinger RN. Traumatic atlanto-occipital dislocation in a child. Spine 1992;S17:1539–1541.
123. Georgopoulos G, Pizzutillo PD, Lee MS. Occipito-atlanto instability in children. A report of five cases and review of the literature. J Bone Joint Surg [Am] 1987;S69:429–436.
124. Lynch JM, Meza MP, Pollack IF, et al. Direct injury to the cervical spine of a child by a lap-shoulder belt resulting in quadriplegia: case report. J Trauma 1996;S41:747–749.
125. Arlet V, Aebi M. Anterior and posterior cervical spine fusion and instrumentation. In: Weinstein SL, ed. Pediatric spine surgery, 2nd ed. Philadelphia: Lippincott Williams & Wilkins, 2001.
126. Heller JG, Jeffords P. Internal fixation of the cervical spine. C. Posterior instrumentation of the lower cervical spine. In: Frymoyer JW, Wiesel SW, eds. The adult and pediatric spine. Philadelphia: Lippincott Williams & Wilkins, 2004.
127. Menezes AH. Surgical approaches to the craniocervical junction. In: Weinstein LD, ed. Pediatric spine surgery, 2nd ed. Philadelphia: Lippincott Williams & Wilkins, 2001.
128. Sponseller PD, Cass J. Atlanto-occipital arthrodesis for instability with neurologic preservation. Spine 1997;S22:344–347.
129. Judd DB, Liem LK, Petermann G. Pediatric atlas fracture: a case of fracture through a synchondrosis and review of the literature. Neurosurgery 2000;S46:991–995.
130. DiBenedetto T, Lee CK. Traumatic atlanto-occipital instability: a case report with follow-up and a new diagnostic technique. Spine 1990;S15:595–597.
131. Koop SE, Winter RB, Lonstein JE. The surgical treatment of instability of the upper part of the cervical spine in children and adolescents. J Bone Joint Surg [Am] 1984;S66:403–411.
132. Wertheim SB, Bohlman HH. Occipitocervical fusion: indications, technique, and long-term results. J Bone Joint Surg [Am] 1987;S69:833–836.
133. Grantham SA, Dick HM, Thompson RC, Stinchfield FE. Occipitocervical arthrodesis. Clin Orthop 1969;65:118–129.
134. Bivins HG, Ford S, Bezmalinovic Z, et al. The effect of axial traction during orotracheal intubation of the trauma victim with an unstable cervical spine. Ann Emerg Med 1988;S17:25–29.
135. Jefferson G. Fracture of the atlas vertebra: report of four cases and a review of those previously recorded. Br J Surg 1919–1920;S7:407–422.
136. Marlin AE, Gayle RW, Lee JF. Jefferson fractures in children. J Neurosurg 1983;S58:277–279.
137. Mikawa Y, Watanabe R, Yamano Y, et al. Fractures through a synchondrosis of the anterior arch of the atlas. J Bone Joint Surg [Br] 1987;S69:483.
138. Richards PG. Stable fractures of the atlas and axis in children. J Neurol Neurosurg Psychiatry 1984;S47:781–783.
139. Tolo VT, Weiland AJ. Unsuspected atlas fractures and instability associated with oropharyngeal injury: case report. J Trauma 1979;S19:278–280.
140. Dvorak J, Panjabi M, Gerber M, et al. CT-functional diagnostics of the rotatory instability of the cervical spine: 1. An experimental study on cadavers. Spine 1987;S12:197–205.
141. Griffiths SC. Fracture of the odontoid process in children. J Pediatr Surg 1972;S7:680–683.
142. Seimon LP. Fracture of the odontoid process in young children. J Bone Joint Surg [Am] 1977;S59:943–948.
143. Odent T, Langlais J, Glorion C, et al. Fractures of the odontoid process: a report of 15 cases in children younger than 6 years. J Pediatr Orthop 1999;S19:51–54.
144. Apple JS, Kirks DR, Merten DF, et al. Cervical spine fractures and dislocations in children. Pediatr Radiol 1987;S17:45–49.
145. Ries MD, Ray S. Posterior displacement of an odontoid fracture in a child. Spine 1986;S11:1043–1044.
146. Scapinelli R. Three-dimensional computed tomography in infantile atlantoaxial rotatory fixation. J Bone Joint Surg [Br] 1994;S76:367–370.
147. Shaw BA, Murphy KM. Displaced odontoid fracture in a 9-month-old child. Am J Emerg Med 1999;S1:73–75.
148. Sherburn EW, Day RA, Kaufman BA, et al. Subdental synchondrosis fracture in children: the value of 3-dimensional computerized tomography. Pediatr Neurosurg 1996;S25:256–259.
149. Stevens JM, Chong WK, Barber C, et al. A new appraisal of abnormalities of the odontoid process associated with atlantoaxial subluxation and neurological disability. Brain 1994;S117:133–148.
150. Godard J, Hadji M, Raul JS. Odontoid fractures in the child with neurologic injury. Direct osteosynthesis with a cortico-spongious screw and literature review. Childs Nerv Syst 1997;S13:105–107.

151. Price E. Fractured odontoid process with anterior dislocation. J Bone Joint Surg [Br] 1960;S42:410–413.
152. Schippers N, Koñigs D, Hassler W, et al. Typical and atypical fractures of the odontoid process in young children. Report of two cases and a review of the literature. Acta Neurochir (Wien) 1996;S138:524–530.
153. Fielding JW. Cineroentgenography of the normal cervical spine. J Bone Joint Surg [Am] 1957;S39:1280–1288.
154. Fielding JW, Griffin PP. Os odontoideum: an acquired lesion. J Bone Joint Surg [Am] 1974;S56:187–190.
155. Fielding JW, Hawkins RJ. Atlanto-axial rotary fixation (fixed rotary subluxation of the atlanto-axial joint). J Bone Joint Surg [Am] 1977;S59:37–44.
156. Fielding JW, Hensinger RN, Hawkins RJ. Os odontoideum. J Bone Joint Surg [Am] 1980;S62:376–383.
157. Fielding JW, Stillwell WT, Chynn KY, et al. Use of computed tomography for the diagnosis of atlanto-axial rotatory fixation. J Bone Joint Surg [Am] 1978;S60:1102–1104.
158. Hawkins RJ, Fielding JW, Thompson WJ. Os odontoideum: congenital or acquired. J Bone Joint Surg [Am] 1976;S58:413.
159. Hukda S, Ora H, Okabe N, Tazima K. Traumatic atlantoaxial dislocation causing os odontoideum in infants. Spine 1980;S5:207–210.
160. Kuhns LR, Loder RT, Farley FA, et al. Nuchal cord changes in children with os odontoideum: evidence for associated trauma. J Pediatr Orthop 1998;18:815–819.
161. Ricciardi JE, Kaufer H, Louis DS. Acquired os odontoideum following acute ligament injury. J Bone Joint Surg [Am] 1976;S58:410–412.
162. Schuler TC, Kurz L, Thompson E, et al. Natural history of os odontoideum. J Pediatr Orthop 1991;S11:222–225.
163. Stillwell WT, Fielding W. Acquired os odontoideum. Clin Orthop 1978;S135:71–73.
164. Verska JM, Anderson PA. Os odontoideum. A case report of one identical twin. Spine 1997;S22:706–709.
165. Giannestras NJ, Mayfield FH, Maurer J. Congenital absence of the odontoid process. J Bone Joint Surg [Am] 1964;S46:839–843.
166. Sherk HH, Dawoud S. Congenital os odontoideum with Klippel-Feil anomaly and fatal atlanto-axial instability. Spine 1981;S6:42–45.
167. Wollin DG. The os odontoideum. J Bone Joint Surg [Am] 1971;S45:1459–1471.
168. Dyck P. Os odontoideum in children: neurological manifestations and surgical management. Neurosurgery 1978;S2:93–99.
169. Brockmeyer DL, York JE, Apfelbaum RI. Anatomic suitability of C1-2 transarticular screw placement in pediatric patients. J Neurosurg 2000;S92(Suppl 1):7–11.
170. Wang J, Vokshoor A, Kim S, et al. Pediatric atlantoaxial instability: management with screw fixation. Pediatr Neurosurg 1999;S30:70–78.
171. Steel HH. Anatomical and mechanical consideration of the atlanto-axial articulation. J Bone Joint Surg [Am] 1968;S50:1481–1482.
172. Lui T-N, Lee S-T, Wong C-W, et al. C1-C2 fracture-dislocations in children and adolescents. J Trauma 1996;S40:408–411.
173. Gallie WE. Fractures and dislocations of the cervical spine. Am J Surg 1939;S46:495–499.
174. Burke SW, French HG, Roberts JM, et al. Chronic atlantoaxial instability in Down syndrome. J Bone Joint Surg [Am] 1985;S67:1356–1360.
175. Dawson EG, Smith L. Atlanto-axial subluxation in children due to vertebral anomalies. J Bone Joint Surg [Am] 1979;S61:582–587.
176. Hammerschlag W, Ziv I, Wald U, et al. Cervical instability in an achondroplastic infant. J Pediatr Orthop 1988;S8:481–484.
177. Kobori M, Takahashi H, Mikawa Y. Atlanto-axial dislocation in Down’s syndrome: report of two cases requiring surgical correction. Spine 1986;S11:195–200.
178. Kransdorf MJ, Wherle PA, Moser RP Jr. Atlantoaxial subluxation in Reiter’s syndrome. Spine 1988;S13:12–14.
179. Miz GS, Engler GL. Atlanto-axial subluxation in Larsen’s syndrome: a case report. Spine 1987;S12:411–412.
180. Sherk HH, Whitaker LA, Pasquariello PS. Fascial malformations and spinal anomalies: a predictable relationship. Spine 1982;S7:526–531.
181. Committee on Sports Medicine. Atlantoaxial instability in Down syndrome. Pediatrics 1984;S74:152–154.
182. Davidson RG. Atlantoaxial instability in individuals with Down syndrome: a fresh look at the evidence. Pediatrics 1988;S81:857–865.
183. Pueschel SM. Atlantoaxial subluxation in Down syndrome. Lancet 1983;S1:980.
184. Pueschel SM, Scolia FH. Atlantoaxial instability in individuals with Down syndrome: epidemiologic, radiographic, and clinical studies. Pediatrics 1987;S4:555–560.
185. Van Dyke DC, Gahagan CA. Down syndrome: cervical spine abnormalities and problems. Clin Pediatr 1988;S27:415–418.
186. Wind WM, Schwend RM, Larson J. Sports for the physically challenged child. J Am Acad Orthop Surg 2004;S12:126–137.
187. Special Olympics Bulletin. Participation by individuals with DS who suffer from atlantoaxial dislocation. Washington, DC: Special Olympics, 1983.
188. Segal LS, Drummond DS, Zanotti RM, et al. Complications of posterior arthrodesis of the cervical spine in patients who have Down syndrome. J Bone Joint Surg [Am] 1991;S73:1547–1560.
189. Windell J, Burke SW. Sports participation of children with Down syndrome. Orthop Clin North Am 2003;S34:439–443.
190. Wetzel FT, Larocca H. Grisel’s syndrome. A review. Clin Orthop 1989;S240:141–152.
191. Kawabe N, Hirotoni H, Tanaka O. Pathomechanism of atlanto-axial rotatory fixation in children. J Pediatr Orthop 1989;S9:569–574.
192. Maheshwaran S, Sgouros S, Jeyapalan K, et al. Imaging of childhood torticollis due to atlanto-axial rotatory fixation. Childs Nerv Syst 1995;S11:667–671.
193. Hohl M, Baker HR. The atlanto-axial joint: roentgenographic and anatomical study of normal and abnormal motion. J Bone Joint Surg [Am] 1964;S46:1739–1752.
194. Geehr RB, Rothman SLG, Kier EL. The role of computed tomography in the evaluation of upper cervical spine pathology. Comput Tomogr 1978;S2:79–97.
195. Pizzutillo PD, Rocha EF, D’Astous J, et al. Bilateral fractures of the pedicle of the second cervical vertebra in the young child. J Bone Joint Surg [Am] 1986;S68:892–896.
196. Phillips WA, Hensinger RN. The management of rotatory atlanto-axial subluxation in children. J Bone Joint Surg [Am] 1989;S71:664–668.
197. Roche CJ, O’Malley M, Dorgan JC, Carty HM. A pictorial review of atlanto-axial rotatory fixation: key points for the radiology. Clin Radiol 2001;S56:947–958.
198. Fardon DF, Fielding JW. Defects of the pedicle and spondylolisthesis of the second cervical vertebra. J Bone Joint Surg [Br] 1981;S63:526–528.
199. Francis WR, Fielding JW, Hawkins RJ, et al. Traumatic spondylolisthesis of the axis. J Bone Joint Surg [Br] 1981;S63:313–318.
200. Howard AW, Letts RM. Cervical spondylolysis in children: is it posttraumatic? J Pediatr Orthop 2000;S20:677–681.
201. Kleinman PK, Shelton YA. Hangman’s fracture in an abused infant: imaging. Pediatr Radiol 1997;S27:776–777.
202. Parisi M, Lieberson R, Shatsky S. Hangman’s fracture or primary spondylolysis: a patient and a brief review. Pediatr Radiol 1991;S21:367–368.
203. Reinges MHT, Mayfrank L, Royhde V, et al. Surgically treated traumatic synchondrotic disruption of the odontoid process in a 15-month-old girl. Childs Nerv Syst 1998;S14:85–87.
204. Ruff SJ, Taylor TKF. Hangman’s fracture in an infant. J Bone Joint Surg [Br] 1986;S68:702–703.
205. Matthews LS, Vetter LW, Tolo VT. Cervical anomaly stimulating hangman’s fracture in a child. J Bone Joint Surg [Am] 1982;S64:299–300.
206. Nordstrom REA, Lahdenrants TV, Kaitila II, et al. Familial spondylolisthesis of the axis is vertebra. J Bone Joint Surg [Br] 1986;S68:704–706.
207. Smith T, Skinner SR, Shonnard NH. Persistent synchondrosis of the second cervical vertebra simulating a hangman’s fracture in a child. J Bone Joint Surg [Am] 1993;S75:1228–1230.
208. Williams JP III, Baker DH, Miller WA. CT appearance of congenital defect resembling the hangman’s fracture. Pediatr Radiol 1999;S29:549–550.
209. Finch GD, Barnes MJ. Major cervical spine injuries in children and adolescents. J Pediatr Orthop 1998;S18:811–814.
210. Jones ET, Hensinger RN. Cervical spine injuries in children. Contemp Orthop 1982;S5:17–23.
211. McClain RF, Clark CR, El-Khoury GY. C6–7 dislocation in a neurologically intact neonate: a case report. Spine 1989;S14:125–126.
212. Schwarz N, Genelin F, Schwarz AF. Post-traumatic cervical kyphosis in children cannot be prevented by nonoperative methods. Injury 1994;S25:173–175.
213. Shacked I, Ram Z, Hadani M. The anterior cervical approach for traumatic injuries to the cervical spine. Clin Orthop 1993;S292:144–150.
214. Hall JE, Denis F, Murray J. Exposure of the upper cervical spine for spinal decompression by a mandible and tongue-splitting approach. Case report. J Bone Joint Surg [Am] 1977;S59:121–125.
215. Magerl F, Seeman P. Stable posterior fusion of the atlas and axis by transarticular screw fixation. In: Kehr P, Weidner A, eds. Cervical spine Vienna: Springer-Verlag, 1985.
216. Roy-Camille R, Saillant G, Mazel C. Internal fixation of the unstable cervical spine by posterior osteosynthesis with plates and screws. In: Sherk HH, ed. The cervical spine, 2nd ed. Philadelphia: JB Lippincott, 1989.