Incidence

The overall incidence of spinal injuries in children in the general population is 7.41 per 100,000. The true incidence may, however, be higher than reported because of failure to recognize these injuries. Aufdermaur found evidence of fractures of the spine at autopsy in 12 of 100 children over an 8-year period. Seven injuries occurred in the cervical spine, four occurred in the thoracic spine, and one occurred in the lumbar spine. Importantly, only one of the 12 subjects had been suspected of having a spinal fracture before autopsy.

Mechanism of Injury

Although these injuries are rare, a high index of suspicion is warranted in polytraumatized children, especially those with head injuries. Approximately 25% to 50% of children with a cervical spine injury have associated head trauma, and as a consequence of this comorbidity, the mortality rate is higher in children with spinal injuries than in their adult counterparts.

In older children, sports injuries, diving accidents, and gunshot injuries are the most common causes. ^∗

∗

The most common mechanisms of injury in young children are motor vehicle accidents, pedestrian–vehicle accidents, falls, or nonaccidental trauma (child abuse). Polk-Williams and colleagues reviewed the National Trauma Data Bank from 2001 to 2005 for patients younger than 3 years that were injured via blunt trauma. The incidence of cervical spine injury was 1.6%, and the most common mechanisms were motor vehicle crashes (66%) or falls (15%).

In infants and young children, nonaccidental trauma is a significant cause of injury to the spine. Nonaccidental trauma was identified in 3.2% of spinal injuries at a level I pediatric trauma center over an 8-year period. The mechanism was nonaccidental trauma in 19% of children 3 years or younger and 38% for those younger than 2 years. These injuries are often associated with other typical stigmata of child abuse including fractures of the skull, ribs, or long bones and cutaneous lesions. The cervical spine is a common location of spinal injuries in abused children (73%), and multilevel injuries are frequent. Upper cervical ligamentous injuries, avulsion fractures of the spinous processes, fractures of the pars or pedicles (most commonly C2), or compression fractures of multiple vertebral bodies are common patterns of injury and are thought to result from severe shaking or battering. Thoracic and lumbar injuries are less common. Displaced fractures through the thoracolumbar neurocentral synchondroses may occur in young children.

In neonates, birth trauma is the most common cause of injury to the cervical spine. Spinal column and spinal cord injuries (SCIs) occur in approximately 1 in 60,000 births and may be an unrecognized cause of death in newborns, as evidenced by autopsy findings of injury to the spinal cord in 10% to 50% of stillborn babies. Excessive distraction and/or hyperextension of the cervical spine is thought to be the most common mechanism of injury and may be associated with abnormal intrauterine position (transverse lie) or a difficult cephalic or breech delivery. When associated with cephalic delivery, the injuries tend to occur in the upper cervical spine and are caused by rotation. Injuries associated with breech delivery are thought to be caused by traction and occur in the lower cervical and thoracic spine. These injuries commonly occur in the absence of osseous injuries. The diagnosis of SCI in neonates is often delayed and should always be considered in a neonate with hypotonia or cardiopulmonary instability or in an older infant with decreased tone, a nonprogressive neurologic deficit, and no history of familial neurologic disorders. A diagnosis can be made with either bedside ultrasound or MRI. Prevention of this injury is preferable. Recognition of intrauterine neck hyperextension in association with breech position may allow for a planned cesarean delivery, which may reduce the risk of SCI.

Initial Evaluation and Transport

Proper care of pediatric spinal injuries begins at the scene of the accident with an appropriate index of suspicion. It should be assumed that a polytraumatized child has a spinal injury until proven otherwise, and appropriate precautions and immobilization should be undertaken. Children should be initially placed in a well-fitting cervical collar and immobilized on a spine board. In the event that commercial adult collars do not fit appropriately, sandbags or towel rolls can be placed on each side of the head to prevent motion. Once the child arrives in the emergency department, every effort should be undertaken to evaluate the child expeditiously. The spine backboard is for transport/transfers only and should be removed from beneath the patient as soon as possible to prevent skin breakdown.

Herzenberg and colleagues were the first to note that transport of young children (younger than 8 years) on a standard adult spine board tended to cause excessive flexion of the cervical spine because of the disproportionately large head diameter relative to the chest in this age group. The obvious concern is that the flexed position of the spinal column could potentially jeopardize the cervical spinal cord, particularly if the mechanism of injury is related to a flexion force, which is often the case in motor vehicle accidents. Therefore, to obtain a neutral position during transport, they recommended using a pediatric spine board with a cutout for the occiput or building up the child’s torso with blankets on a standard spine board. Alternatively, a standard spine board can be used if a towel roll is placed under the shoulders to allow the head to drop into slight extension ( Fig. 12-3 ).

Schematics of two types of spine boards modified for transportation of the young child with a suspected cervical spine injury. Note the occipital recess in the top drawing and the extra padding to elevate the torso in the lower to prevent flexion of the spine by the child’s head, which is disproportionately larger than the chest in young children.

(Reproduced from Dormans JP: Evaluation of children with suspected cervical spine injury. Instr Course Lect 51:403, 2002.)

In a subsequent study, Curran and colleagues prospectively evaluated methods of positioning the child to achieve neutral alignment of the cervical spine after trauma. They measured sagittal alignment on supine lateral radiographs in 118 pediatric trauma patients and determined that only 60% were within 5° of neutral alignment. They suggested that younger children might need more relative chest elevation to avoid flexion of the head and cervical kyphosis after immobilization. These findings were confirmed by Nypaver and Treloar, who determined in their study that children younger than 4 years required an additional 5 mm of elevation of the torso, on average, than those older than 4 years to achieve neutral cervical spine alignment.

Although it is not known how alignment of the cervical spine during transport affects outcome, it seems prudent to avoid flexion of the neck by following the recommendations of Herzenberg and associates regarding spine board immobilization, keeping in mind that the very young may need additional elevation to achieve neutral alignment. A pediatric-sized cervical collar and appropriate positioning may not be enough to ensure neutral alignment of the cervical spine in young children. As a practical guideline for proper positioning of a child on the spine board during transport, the external auditory meatus should be aligned with or slightly posterior to the shoulders.

Cervical Spine Considerations

The likelihood of missing a cervical spine injury has been reported to be increased almost twenty-threefold in children who are incapable of verbal communication for whatever reason; therefore a thorough, careful examination is essential when a polytraumatized child is evaluated. Historically, several mechanisms of injury are considered to be risk factors for overt or occult injury to the cervical spine: falls from a distance greater than the height of the child, pedestrian– or cyclist–motor vehicle accidents, and motor vehicle accidents with unrestrained occupants. Head or facial trauma, altered mental status, and/or loss of consciousness are also considered to be risk factors. Neck pain, guarding, and torticollis are the most reliable signs of an injury to the cervical spine in children. Extremity weakness, sensory changes (numbness or tingling), bowel and bladder dysfunction, and, less frequently, headaches, seizures, syncope, and respiratory distress are signs heralding injury to the spinal cord ( Table 12-1 ). If any of these conditions are present, immobilization of the cervical spine should be continued or initiated until imaging studies can be completed and the spine can be cleared.

The physician should palpate the entire spine, from occiput to the sacrum, while keeping in mind that spinal injuries can occur at multiple levels and are noncontiguous in up to 38%. Approximately 16% of noncontiguous spinal injuries may be initially overlooked; thus one must maintain a high index of suspicious for other injuries. During the examination, the cervical collar should be carefully removed, and an assistant should stabilize the head so that the patient or examiner does not inadvertently move the head. Cervical spine injuries often are seen with torticollis, so it is important to note the position of the head and the presence of asymmetry in alignment. The anterior and posterior sides of the neck are examined for lacerations and wounds. The cervical spine is palpated anteriorly and posteriorly for the presence of tenderness or interspinous widening. The collar is reapplied, and attention is directed to the thoracolumbar spine.

Thoracolumbar Spine Considerations

The possibility of injury to the thoracolumbar spine should always be suspected in children who are comatose, have a distracting injury, or are not verbal due to age. The thorax and abdomen need to be inspected for signs of trauma. Abdominal injuries, particularly those of the small bowel, are associated with flexion–distraction injuries of the thoracolumbar spine and are often heralded by the presence of contusions or abrasions caused by lap belts. The patient should be logrolled while the head and neck are kept in alignment with the rest of the spine and trunk. During this maneuver, cervical in-line traction should be avoided, particularly in young children, because of the increased risk of ligamentous and atlantooccipital injuries. The thoracolumbar spine should be palpated along the spinous processes for evidence of tenderness, interspinous widening, or malalignment. The finding on physical examination of midline or paravertebral pain alone is predictive of the presence of a fracture of the thoracolumbar spine with a sensitivity of 87% and a specificity of 75%. Evaluation of the entire thoracolumbar spine is critical; otherwise, a noncontiguous fracture may be missed. A thorough neurological examination is also important given that 20% to 30% will have a deficit.

Neurologic Examination

For all suspected spinal injuries, an accurate baseline neurologic examination should be carefully documented in patients who are conscious and cooperative. The sensory examination should include evaluation of light touch, pain, and proprioceptive function. Pain and temperature sensation are mediated by the spinothalamic tract that traverses the anterolateral column of the spinal cord. This can be assessed with the use of a clean needle to test pinprick sensation and an alcohol pad for temperature discrimination. Light touch and proprioceptive (position) sensation are functions of the posterior spinal columns. One can test light touch by stroking the extremity with a piece of paper, and one can test proprioception by asking the patient to determine directional change in the position of a finger or toe.

Dermatomal patterns of sensation correlate with the spinal nerve roots exiting specific anatomic levels of the spinal cord ( Fig. 12-4 ). C1 and C2 innervate the occipital region, C3 and C4 innervate the nape of the neck, C5 innervates the deltoid region, C6 innervates the radial aspect of the forearm, C7 innervates the long finger, C8 innervates the ulnar border of the hand, and T1 innervates the medial border of the arm. The chest and abdomen are innervated by the T2–T12 nerve roots. Specifically, T4 provides sensation at the nipple line, T10 provides sensation at the umbilicus, and T12 provides sensation at the inguinal ligament. In the lower extremities, the pattern of sensory innervation mirrors the embryonic rotational maturation of the limbs. L1 and L2 contribute innervation below the inguinal ligament to the medial thigh, L3 provides sensation to the anterior midthigh, L4 provides sensation to the knee region and medial calf, L5 provides sensation to the lateral calf and first web space, and S1 provides sensation to the lateral aspect and sole of the foot. The perineal region is innervated by the S3–S5 roots. Preservation of function at this level, referred to as sacral sparing, is important because it indicates that some of the spinal tracts are still intact and that the injury to the spinal cord is incomplete; therefore it is associated with a better prognosis for neurologic recovery.

Schematic of sensory dermatomes.

(From Keenen TL, Benson DR: Initial evaluation of the spine-injured patient. In Browner BD, Jupiter JB, Levine AM, et al, editors: Skeletal trauma: fractures, dislocations, ligamentous injuries, vol 1, Philadelphia, 1992, W.B. Saunders, p 594.)

Motor function should be graded on a scale of 0 to 5 with grade 0 indicating complete paralysis, grade 1 indicating trace function, grade 2 indicating full range of joint motion with gravity eliminated, grade 3 indicating antigravity function, grade 4 indicating function against slight resistance, and grade 5 indicating normal strength against resistance. The level of SCI can be assessed by the presence or absence of function in key muscle groups . In the upper extremities, C5 innervates the muscles responsible for elbow flexion; C6, wrist extension; C7, wrist flexion; C8, finger flexion; and T1, finger abduction. In the lower extremities, L2 innervates hip flexion, L3, knee extension; L4, ankle dorsiflexion; L5, great toe extension; and S1, ankle plantar flexion.

Deep tendon reflexes should be graded as absent (0), hypoactive (1), normal (2), or hyperreflexic (3). In the upper extremities, the biceps tendon reflex is mediated by the C5 nerve root, the brachioradialis is mediated by C6, and the triceps is mediated by C7. In the lower extremity, the patellar tendon reflex is mediated by L4, and the Achilles tendon is mediated by S1.

The abdominal, Babinski, and bulbocavernosus reflexes should also be assessed. The abdominal reflex is performed by division of the belly into four quadrants, with the umbilicus at the center. When the skin in each of the quadrants is stroked, the umbilicus should deviate in that direction. Absence of a response may signify an upper motor neuron lesion, whereas asymmetric loss of the reflex may indicate a localized lower motor neuron lesion. The Babinski test is performed by stroking the lateral plantar aspect of the foot. A pathologic response is indicated by an upgoing great toe and is indicative of an upper motor neuron lesion.

The bulbocavernosus reflex is an important test for determination of the status of an injury to the SCI. The test involves a digital rectal examination with simultaneous application of traction on an in-dwelling Foley catheter (or with squeezing of the glans penis or clitoris). The presence of the reflex is indicated by concurrent contraction of the anal sphincter and heralds the end of spinal shock. Spinal shock is a transient phenomenon that occurs within the first 24 hours of SCI and is thought to be due to swelling about the neural structures within the spinal column. Once spinal shock has passed, as indicated by return of the bulbocavernosus reflex, the status of the SCI can be predictably characterized. This reflex is less reliable with injuries around the conus medullaris (T12–L2) because the afferent nerve fibers that mediate the reflex lie within the zone of injury and may be directly affected. As a consequence, return of the reflex may take much longer in this group of patients. Additionally, all patients with a significant spinal injury should have an evaluation of bladder function with the use of postvoiding straight catheterization.

When an accurate neurologic examination cannot be obtained because of the child’s age or altered mental status, findings that may suggest an SCI in the initial evaluation period include flaccidity, diaphragmatic breathing without the assistance of accessory muscles, priapism, and the presence of clonus. Evaluation of a patient in this setting should include inspection and palpation of the spine from the occiput to the sacrum, assessment of motor and sensory function as determined by the ability to withdraw from painful stimuli, and testing of deep tendon, abdominal, Babinski, and bulbocavernosus reflexes.

Indications

The National Emergency X-Radiography Utilization Study (NEXUS) is a decision-making instrument that has been used in adults to determine the need for radiographic imaging of the cervical spine after trauma. The criteria for clinical clearance are absence of neck pain/midline cervical tenderness, neurologic symptoms, distracting injuries, or altered mental status (due to injury or intoxication). If any of these conditions are present, the patient is considered to be at high risk of a spinal injury and must be evaluated with radiographs. Application of this protocol in the pediatric population was studied by Viccellio and colleagues in a prospective, multicenter study of 3065 patients who were evaluated with the use of the NEXUS instrument before undergoing radiographic imaging. All of those placed in the high-risk group underwent anteroposterior (AP), lateral, and open-mouth odontoid radiographs. The instrument correctly placed all 30 cervical spine injuries into a high-risk group, and imaging confirmed the presence of an injury in each instance. More importantly, no cervical spine injuries were noted in the low-risk group, which yielded a negative predictive value of 100%. One fault of this study is that only four of the injured children were younger than 9 years, and none was younger than 2 years. A second concern is the possibility of false-negative radiographs in the low-risk group, given the recognized limitations of plain radiography in detecting injuries of the cervical spine. The authors concluded that application of the NEXUS criteria in an appropriate age group could potentially decrease pediatric cervical spine imaging by nearly 20%. They also cautioned that NEXUS rules should not be applied if the child is very young, if an accurate history and examination cannot be obtained, or if associated injuries heighten the suspicion of a spinal injury.

Garton and Hammer retrospectively evaluated the NEXUS criteria in 190 pediatric patients with documented cervical spine injuries. Using the NEXUS criteria to determine the need for cervical imaging would have resulted in no missed injuries among the 157 patients older than 8 years. However, in the 33 patients younger than 8 years, two injuries would have been missed (94% sensitivity). The two missed injuries were in the upper cervical region and in patients younger than 2 years.

The Canadian C-spine (cervical spine) Rule (CCR) was designed to guide the decision for cervical imaging in adult trauma patients. Imaging is recommended if the patient is high risk (65 years or older, dangerous mechanism, or extremity paresthesias) and if the patient is unable to actively rotate the neck to 45° in each direction (tested only in the presence of low-risk factors). The CCR and NEXUS criteria were retrospectively evaluated by Ehrlich and colleagues in trauma patients younger than 10 years. Both criteria would have missed important cervical injuries. The CCR had a sensitivity of 86% and a specificity of 94%, whereas the NEXUS had a sensitivity of 43% and a specificity of 96%. The authors concluded that these criteria are not adequate for the pediatric population as currently designed.

Laham and colleagues defined children at high risk of cervical spine injury as those who were incapable of verbal communication because of young age (<2 years old), those with altered mental status, and those with neck pain. They retrospectively evaluated 268 children with isolated head injuries and, using these criteria, placed 133 in the high-risk group and 135 in the low-risk group. They identified fractures in 10 children in the high-risk group (7.5%) and no fractures in the low-risk group.

In a multicenter review, Pieretti-Vanmarcke and associates evaluated 12,537 blunt trauma patients younger than 3 years and created a scoring system using four independent predictors of cervical spine injury: Glasgow Coma Scale (GCS) score of less than 14 (3 points), GCS eye score of 1 (2 points), motor vehicle accident mechanism (2 points), and age 25 to 36 months (1 point). A total score of 0 or 1 had a negative predictive value of 99.93% for cervical spine injury. The five outliers with injuries despite scores of 0 or 1 had associated facial or skull fractures, loss of consciousness, or neck splinting.

Imaging of the cervical spine after trauma should be undertaken if the mechanism of injury is high risk and if the child is nonverbal because of age or altered mental status, is intoxicated, has a neurologic deficit (persistent or transient), complains of neck pain, exhibits physical signs of neck or lap belt trauma, or has sustained other painful distracting injuries ( Table 12-1 ). Unexplained cardiorespiratory instability can be an indication of a high cervical spine injury and should be evaluated appropriately. Imaging is not required for children who are communicative, alert, and nonintoxicated and have no neck pain, neurologic deficit (transient or persistent), mental status change, or painful distracting injury. There is a paucity of literature on clinical clearance of the thoracolumbar spine. However, if a fracture is found at one level of the spine, the remaining spine should be imaged because of the high risk of a noncontiguous injury.

Plain Radiography, Cervical

Plain radiography of the cervical spine has been studied most extensively because of the length of time it has been available. A single supine lateral cervical radiograph with visualization of all seven cervical vertebrae, including the occipitocervical and cervicothoracic junctions, has a reported sensitivity of approximately 80% in the pediatric population. Lally and colleagues found that all seven cervical vertebrae were seen in only 57% of children on the initial cervical spine series, usually because of difficulty in visualizing the cervicothoracic junction. Visualization of the cervicothoracic junction can be improved with traction or the so-called swimmer’s view, in which the arm is extended overhead. The addition of AP and open-mouth odontoid radiographs increases the sensitivity of plain radiography to approximately 94% if adequate images can be obtained. However, the open-mouth odontoid can be especially challenging to obtain in the young child. Buhs and colleagues performed a multiinstitutional retrospective review on children younger than 16 years with documented cervical spine injuries. Standard AP and lateral radiographs confirmed the diagnosis in 13 of 15 children younger than 9 years. In none of the 15 patients did the open-mouth odontoid view provide any additional information. In only one of 36 patients 9 to 16 years of age was the open-mouth odontoid deemed to be beneficial: it identified a type III odontoid fracture. The authors concluded that the open-mouth odontoid view was not helpful in children younger than 9 years. Instead, they recommended the use of computed tomography (CT) to evaluate the upper cervical spine from the occiput to C2. Similarly, Garton and Hammer noted that, for children younger than 8 years, the use of CT (occiput to C3) in combination with radiographs was more sensitive (94%) than radiographs with flexion–extension views (81%), or radiographs alone (75%).

Normal Radiographic Alignment

Interpretation of cervical spine radiographs in children requires an understanding of normal anatomy as well as the normal anatomic variants of the immature spine that can sometimes mimic trauma. On the lateral radiograph, the vertebral bodies, lamina, and spinous processes should be aligned in a gentle lordotic contour, and the facet joints should overlap symmetrically ( Fig. 12-5 ). Assessment should proceed methodically in a cephalic to caudal direction so that no injuries are missed. Special attention should be given to evaluating the upper cervical (atlantoaxial) and craniocervical (atlantooccipital) region given the propensity for injuries to this area in children and the subtlety of associated radiographic findings.

Normal relationships seen on the lateral cervical spine radiograph: 1, spine processes; 2, spinolaminar line; 3, posterior vertebral body line; 4, anterior vertebral body line.

(Reproduced from Copley LA, Dormans JP: Cervical spine disorders in infants and children. J Am Acad Orthop Surg 6:205, 1998.)

Several methods for evaluating alignment at the craniocervical junction with the use of the lateral radiograph have been described. The Wackenheim line is drawn along the posterior aspect of the clivus toward the odontoid process and should intersect the superior/posterior aspect of the odontoid process ( Fig. 12-6 ). An anterior or posterior shift of this line relative to the odontoid indicates anterior or posterior displacement, respectively, of the occiput on the atlas. In children younger than 13 years the odontoid may not be completely ossified, giving the false impression that an anterior atlantooccipital dislocation is present. Powers ratio is determined by a line drawn from the basion to the posterior arch of the atlas and a second line drawn from the opisthion to the anterior arch of the atlas ( Fig. 12-7 ). Ratios of more than 1.0 and less than 0.55 represent anterior and posterior displacements of the atlantooccipital joint, respectively. However, the ratio is much more sensitive for anterior dislocations. The Rule of 12s described by Harris is another method for evaluating instability between the occiput and C1. In this technique, the distance between the basion and tip of the odontoid process (basion–dens interval [BDI]) should be 12 mm or less, and a line drawn superiorly from the posterior aspect of the body of C2 and odontoid (basion–axis interval [BAI]) should pass within 12 mm of the basion ( Fig. 12-8 ). The BDI is unreliable in children younger than 13 years old because of incomplete ossification of the odontoid; however, the BAI should still be less than 12 mm. Distraction at the atlantooccipital junction can be detected by measurement of the vertical height of the atlantooccipital junction. According to Kaufman and colleagues, this distance should not exceed 5 mm at any point in the normal spine ( Fig. 12-9 ).

Wackenheim line is drawn tangentially to the clivus and should intersect the superior–posterior aspect of the odontoid process. A shift of the intersection point anteriorly or posteriorly is indicative of atlantooccipital displacement in the same direction. In the young child, incomplete ossification of the odontoid can give the false impression that there is an anterior translation.

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 line BC is divided by the length of the line OA. A ratio of more than 1 is diagnostic of anterior atlantooccipital translation, and a ratio of less than 0.55 is less sensitive for posterior translation.

(From Hosalkar HS, Cain E, Horn D, et al: Traumatic atlanto-occipital dislocation in children. J Bone Joint Surg Am 87:2480–2488, 2005.)

The Harris method. The distance between the basion and the tip of the odontoid (basion–dens interval [BDI] ) should be ≤ 12 mm, and the distance from the dens to a line drawn superiorly from the posterior aspect of the body of C2 along the odontoid (basion–axis interval [BAI] ) should also be less than 12 mm.

(From Hosalkar HS, Cain E, Horn D, et al: Traumatic atlanto-occipital dislocation in children. J Bone Joint Surg Am 87:2480–2488, 2005.)

The Kaufman method. The width of the atlantooccipital junction including the distance between the occipital condyles and articular surfaces of the lateral masses of C1 (facet–condylar distance) should not exceed 5 mm at any point.

(From Hosalkar HS, Cain E, Horn D, et al: Traumatic atlanto-occipital dislocation in children. J Bone Joint Surg Am 87:2480–2488, 2005.)

Integrity of the atlantoaxial articulation can be evaluated with the use of the atlantodens interval (ADI). The ADI is measured from the posterior aspect of the anterior ring of C1 to the anterior aspect of the odontoid. The normal distance is less than 3 mm in adults, but in children, the normal ADI can be up to 5 mm. If the ADI exceeds 5 mm on lateral flexion and 4 mm on lateral extension, the transverse atlantal ligament (TAL) is likely to be incompetent. When the ADI exceeds 10 to 12 mm, the alar and apical ligaments are probably also insufficient, and the risk of spinal cord compression from instability of C1 on C2 is high. The risk of spinal cord compression due to atlantoaxial instability can also be determined by direct measurement of the space available for the spinal cord (SAC). Steel’s ³⁰³ “rule of thirds” states that, at the level of the dens, one third of the anterior-to-posterior diameter of the spinal column extending from inside the anterior ring of C1 to the inside of the ring posteriorly should be occupied by the odontoid, one third should be occupied by the spinal cord, and one third should be occupied by the SAC. When the ring of C1 shifts anteriorly or the dens migrates posteriorly so that the SAC is reduced to less than one third, the spinal cord is likely to be compressed.

Swischuk described the posterior cervical or spinolaminar line to help diagnose pathologic angulation and translation in the upper cervical spine. This line is drawn from the anterior aspect of the spinous process of C1 to the anterior aspect of the spinous process of C3 and should pass within 1.5 mm of the anterior aspect of the spinous process of C2. If the distance exceeds 1.5 mm, an injury should be suspected ( Fig. 12-10 ).

Swischuk described the posterior cervical line (spinolaminar line), which is drawn from the anterior aspect of the spinous process of C1 to the anterior aspect of the spinous process of C3. If the line does not pass within 1.5 mm of the anterior aspect of the spinous process of C2 on flexion–extension radiographs, a true injury should be suspected.

Plain Radiographs, Thoracolumbar Spine

Conventional radiography is commonly used for screening of the thoracolumbar spine. However, the efficacy of this modality is sometimes hindered by the inability to obtain satisfactory views of the upper two or three thoracic vertebrae. The swimmer’s view can be helpful in such situations to counter obstruction by the shoulders and improve visualization of this transitional region of the spine. Screening radiographs should be obtained as soon as the initial trauma survey is completed so that the process of clearing the thoracolumbar spine is expedited. Radiographs should be evaluated for the presence of fractures, overall alignment of the spine, facet joint symmetry, and interspinous distance. The lines described for evaluating the cervical spine should have a smooth contour in the thoracolumbar spine. Although burst fractures are best seen on CT scans, subtle findings on conventional radiography include interpedicular widening on the AP projection and small cortical defects at the posterior–superior corner of the vertebral body on the lateral projection.

Computed Tomography, Cervical

Inadequate radiographs were the leading cause of missed injuries and subsequent neurologic deterioration in a large series of trauma patients. Having to repeat radiographs when initial studies are suboptimal is also costly and inefficient patient care. Given these deficiencies of conventional radiographs and the continuing advances in digital cross-sectional imaging, including the speed of data acquisition, image resolution, and reformatting capabilities, the use of CT for evaluation of spinal trauma in children has continued to evolve. In adults, the use of CT has replaced conventional radiography as the screening tool of choice for the cervical spine in the setting of blunt trauma. The role of CT in evaluating spine trauma in the pediatric population is still being defined. Keenan and colleagues demonstrated that CT of the cervical spine may be especially beneficial in a subgroup of children who were older than 8 years, unrestrained in a motor vehicle accident, with a GCS score of greater than 13, and intubated. They found that a CT scan obtained early in the evaluation process prevented the need for obtaining multiple equivocal radiographs. Link and colleagues reviewed a series of 202 patients with severe cranial trauma in whom CT scans of the head as well as plain radiographs were performed. Twenty-eight had fractures of C1 or C2, and plain radiographs failed to identify fractures in 11 of the 28. Another 11 had occipital condoylar fractures that were identified only with CT. The authors concluded that routine CT of the craniocervical junction in patients with head trauma is useful for detecting occult fractures of C1, C2, or the occipital condyles.

The risk of developing thyroid cancer from exposure to ionizing radiation is a valid concern. Previous estimates have put the radiation dose associated with CT at up to four times that of conventional radiography. Recently, Muchow and colleagues estimated the radiation exposure of pediatric trauma patients for cervical imaging with plain radiographs (i.e., AP, lateral, and open-mouth odontoid) and multidetector CT (MDCT) by calculating absorbed doses based on imaging protocols. The mean absorbed thyroid dose was approximately 0.9 mGy for radiographs and 64 mGy for CT. With this radiation exposure, the median excess relative risk of thyroid cancer induction with CT was 13% in males and 25% in females compared with 0.24% in males and 0.51% in females for radiographs. This translates to an increase in the absolute risk of thyroid cancer after CT from 5.2 to 5.87/100,000 in males and 15.2 to 19.0/100,000 in females. Actual radiation exposure from a CT scan is dependent on the specific scanner technology and imaging protocols. At the authors’ institution, they have calculated absorbed thyroid doses and whole body effective doses using 64-slice MDCT with age-specific protocols and found the doses to be only two to three times greater than five cervical spine digital radiographs (i.e., AP, lateral, obliques, and open-mouth odontoid). The authors anticipate that further advances in technology will allow the radiation doses from CT to approximate digital radiography.

The central issue is whether the specificity and sensitivity of CT is sufficiently superior to plain radiographs to justify the higher costs and to offset the risks of increased radiation exposure in children. Rana and colleagues reviewed 318 pediatric trauma patients in whom 27 cervical injuries were identified. There were five false-negative results and five false-positive results with radiographs. The sensitivity and specificity of CT were 100% and 98%, respectively, whereas the sensitivity and specificity of plain radiographs were 62% and 1.6%, respectively. Carlan and colleagues retrospectively studied 413 pediatric trauma patients (mean age, 10.7 years) who were evaluated with both conventional radiography and CT. The accuracy (i.e., sensitivity, specificity, and negative predictive value) of CT was determined to be superior to plain radiography as a screening test. Radiographs missed 44% of cervical spine injuries in children younger than age 14 years. They also found that radiographs added no additional diagnostic information compared with CT.

The utility of CT to detect ligamentous injuries when compared with MRI has a reported sensitivity of 17%, a specificity of 100%, a positive predictive value of 100%, and a negative predictive value of 94%. However, a recent study by Gargas and colleagues suggests that technologic advances allowing greater CT scan resolution may make MRI and CT comparable for the detection of significant ligamentous injuries that requires operative treatment. They reviewed 173 pediatric trauma patients with normal cervical CT findings who subsequently had MRI scans. MRI detected five unstable injuries that required operative stabilization in the low-resolution (single-slice) CT group (85 patients); no missed unstable injuries (0 of 88) were found in the high-resolution (64-slice) CT group (88 patients).

On the basis of the available data, either conventional radiography or CT may be used effectively to screen for cervical spine injuries. In children younger than 8 years, a soft tissue or chondral injury is more likely than a fracture, and MRI is arguably a more appropriate study, although it may be logistically difficult to use as a screening study. CT with sagittal and coronal reconstructions is the authors’ preferred method for initially evaluating the cervical spine in children of all ages in the polytrauma setting. When the CT result is normal but concern for an unstable spinal injury remains, the authors use MRI to evaluate for soft tissue or chondral injuries, particularly in children younger than 8 years.

Computed Tomography, Thoracolumbar

Growing evidence is showing that CT is superior to plain radiography for evaluating the thoracolumbar spine in adults. In polytraumatized patients, the use of data reformatted from “traumagrams,” which are CT scans of the head to the pelvis for evaluation of the thoracic and abdominal cavities, has been shown to be very sensitive at screening for injuries of the thoracic, lumbar, and sacral (TLS) spine and, in many centers, has eliminated the need for conventional radiographs. At least five prospective studies have demonstrated a higher sensitivity for CT (93% to 100%) when compared with plain radiography (33% to 74%) and better interobserver reliability. This information can likely be extrapolated to the pediatric population, but further studies need to be performed before CT with sagittal and coronal reconstructions replace conventional radiography for TLS clearance in children.

Magnetic Resonance Imaging

MRI is the study of choice for evaluation of the spinal cord and soft tissue structures, including ligaments, cartilage, and intervertebral disks. It is useful for assisting with cervical spine clearance in the obtunded patient. Young children usually require sedation, so the logistics of performing this study can be cumbersome unless the child is already intubated or obtunded. MRI should be obtained in those with evidence of a neurologic deficit. MRI should also be performed if the neurologic deficit is transient because this may herald an underlying ligamentous injury that requires immobilization.

Recognizing that spinal injuries in young children are more likely to involve the soft tissues or chondral structures, Flynn and colleagues have advocated for the definitive role of MRI, in addition to plain radiographs and CT, in their evaluation protocol. In their study of 74 children, MRI altered the diagnosis in 34%, identifying injuries in 15 patients with normal radiographic results and excluding injuries suspected on plain radiographs and CT scans in seven and two patients, respectively. In 25 obtunded or uncooperative children, MRI demonstrated three with significant injuries. Keiper and colleagues obtained MRI studies in 52 children who had clinical findings that were consistent with a cervical spine injury, despite having negative results on plain radiography and CT. MRI findings were abnormal in 16 of 52 patients, and posterior soft tissue injury was the most common pathology. Henry and colleagues compared MRI with CT for the detection of osseous and ligamentous/soft tissue injuries in 84 pediatric trauma patients. MRI identified all six fractures among these patients as well as an additional compression fracture, whereas CT detected only one of the six ligamentous injuries seen with MRI. If CT is used as the comparison standard for osseous injury, MRI had a sensitivity of 100%, a specificity of 97%, a positive predictive value of 100%, and a negative predictive value of 75%.

MRI can be especially helpful in clearing the cervical spine of possible soft tissue injury in obtunded patients in whom the physical examination may be unreliable or unobtainable. Frank and colleagues evaluated the effectiveness of MRI at decreasing time to cervical spine clearance, length of time in the pediatric ICU, and length of time in the hospital and found a statistically significant decrease for all of these outcome measures. In a study of the use of MRI and CT in a clearance protocol for the obtunded trauma patient, Stassen and associates found a significant number of ligamentous cervical spine injuries (25%) identified by MRI that were missed by CT alone.

In the thoracolumbar spine, MRI also affords effective visualization of injury to the ligaments, intervertebral disks, bones, and spinal cord. Sledge and colleagues found that MRI was capable of determining the fracture classification, spinal stability, and appropriate treatment in 19 children with thoracolumbar fractures and neurologic deficits. They also found that the pattern of early cord changes seen on MRI were predictive of the potential for neurologic recovery.

Spine Clearance Protocols

Cervical

Once a cervical collar has been placed on a child or the neck immobilized, either at the scene of an accident or in the emergency department, formal clearance of the cervical spine is necessary before immobilization may be discontinued. Early cervical spine clearance has multiple benefits and helps avoid known complications including skin breakdown, dysphagia, pulmonary complications, and increased intracranial pressure. No universally accepted protocols for clearing the cervical spine exist for children younger than 8 to 10 years. Hartley and colleagues have proposed an algorithm that permits clearance of the cervical spine based on clinical examination alone if the child is awake, alert, and cooperative, if there are no signs of cervical injury, and if the mechanism of injury is not consistent with cervical trauma. For children who are obtunded or otherwise unable to be examined and all those having a history or findings suggestive of injury to the cervical spine, clearance is based on a five-view cervical spine radiographic series, consisting of AP, lateral, open-mouth odontoid, and oblique views plus CT of the axial region of the spine, from the occiput to C3. The rationale for CT includes the preponderance of injuries in the upper cervical region in children younger than 8 years and the technical difficulty of imaging this area with plain radiographs. In this study, eight of 112 children were given diagnoses of cervical spine injuries. Two of six children with osseous injuries (33%) received diagnoses only by CT scan. No injuries were missed, and cervical immobilization was discontinued in a timely fashion. Alternatively, as proposed by Carlan and colleagues, a CT scan with sagittal and coronal reconstructions can be used in place of plain radiography to clear the cervical spine.

An efficient multidisciplinary approach can facilitate rapid clearance of the cervical spine, decreasing the average time to less than 8 hours in the nonintubated patient and to 20 hours in the intubated child. The ability to rapidly clear the spine is dependent on a protocol that is safe and user-friendly enough for the primary team to perform. Anderson and colleagues have demonstrated the safety and effectiveness of such protocols, reporting a 60% increase in the number of spines appropriately cleared (no late injuries detected) by nonspine physicians after institution of a spinal clearance protocol.

A spinal clearance protocol should incorporate a thorough history and physical examination with judicious use of imaging modalities ( Fig. 12-11 ). Clinical clearance can be performed when the following criteria are met: the child is old enough to effectively communicate pain; the child is fully alert without evidence of intoxication or mental status change; no paraspinal or midline cervical tenderness is present; no evidence of a neurologic deficit or a history of a transient deficit exists; and no distracting injuries are present. If no abnormalities exist, the spine may be cleared without imaging if painless, full range of motion can be achieved without the collar. However, if any of the above criteria are not met or if provocative movement causes pain, then spinal precautions should be continued and imaging should be performed.

Algorithm for cervical spine clearance. GCS, Glasgow Coma Scale.

At a minimum, initial radiographic evaluation should include cross-table lateral and AP views. On the lateral view, it is essential to see the C7–T1 disk space. Open-mouth odontoid radiographs may be considered in children who have the ability to cooperate (generally older than 8 years). Otherwise a CT scan from the occiput to C3 should be performed. Oblique radiographs are additionally helpful in defining details of the pedicles and facet joints. CT provides excellent definition of obvious pathology, confirmatory data of areas suggestive of pathology, and excellent visualization of the occipitocervical and cervicothoracic levels of the cervical spine, which can be difficult to adequately image with plain radiography.

Children undergoing CT scans evaluating possible head injuries should have the cervical spine included in the study. As discussed previously, a CT scan with sagittal and coronal reconstructions can be used in place of plain radiography to effectively evaluate the cervical spine. Imaging studies must adequately visualize the entire cervical spine, from the occipitocervical junction to the T1 vertebral body. Children without evidence of injury on plain radiography or CT scans, but with persistent neck pain or guarding, may undergo MRI for evaluation of soft tissue injury. If none exists, immobilization can be discontinued. An alternative approach for children with neck pain but normal radiographic or CT results is as follows: MRI is rejected in favor of immobilization of the child in a rigid collar for 10 to 14 days so that paraspinal muscle spasm can resolve; flexion–extension radiographs are then performed to exclude instability. In the presence of a neurologic deficit or a history of a transient deficit, an MRI of the entire spine should be obtained. If a spinal cord injury without radiographic abnormality (SCIWORA) is suspected, then the cervical collar should remain in place, spine precautions should be continued, and the patient should be admitted for observation given the risk of delayed neurologic deterioration after these injuries.

Clearing the cervical spine in an unconscious or obtunded patient can also be facilitated by a standard protocol ( Fig. 12-11 ). These patients often undergo a CT traumagram as part of their initial workup to evaluate for associated injuries, at which time the cervical spine can be imaged with sagittal and coronal reconstructions. When the patient’s mental status returns to baseline, the protocol for a conscious patient can be used. If the child has a more significant head injury and is expected to remain unconscious for an extended period of time, MRI can be used to facilitate clearance of the cervical spine so that problems associated with cervical collar immobilization may be avoided. In these patients expeditious use of MRI has also been shown to decrease the time spent in the ICU as well as the overall hospital length of stay. For these reasons, MRI should be obtained as early as is reasonably possible. If the results of plain radiographs or CT with sagittal and coronal reconstructions and the MRI are normal, the cervical spine can be safely cleared.

Characteristics of Spinal Cord Injury

The annual incidence of SCI in the pediatric population in the United States is 19.9 injuries per one million children, which represents approximately 1400 new cases per year. In most studies of pediatric spinal trauma, a spinal fracture without SCI is slightly more common than a fracture with SCI. Vitale and colleagues reviewed the Kid’s Inpatient Database (KID) and the National Trauma Data Bank (NTDB) from 1997 to 2000 and noted that SCIs in children most often occurred as a result of motor vehicle accidents (56%), followed by accidental falls (14%), firearm injuries (9%), and sports (7%). The incidence of SCI was more than twice as common in males. Additionally, alcohol or drugs were involved in 30% of pediatric SCIs, and 67% of those injured in motor vehicle accidents were not restrained. Herndon reported diving accidents as causative of 9% of pediatric SCIs.

In children younger than 10 years, the most common causes are pedestrian–motor vehicle accidents and falls; in children older than 10 years, the most common etiologies are passenger-related motor vehicle accidents, diving, and other sports-related injuries (e.g., gymnastics, diving, downhill skiing, and contact sports). There is a 5% to 10% mortality rate during the first year after SCI in children, and children younger than 11 years are five times more likely to die within the first year after sustaining an injury to the spinal cord.

Because of the anatomic and biomechanical characteristics of the immature spine, SCIs in children younger than 8 years usually involve the upper cervical spine, often with no discernable bony injury, and are twice as likely to result in quadriplegia. Children older than 8 years are more likely to have fractures and injury patterns that more closely mirror the injury pattern in adults. Overall, Rang found that paraplegia was three times more common than quadriplegia in his review of spinal trauma in children at Toronto Sick Children’s Hospital over a 15-year interval. In comparison with adults, children are more likely to have multilevel injuries; approximately 25% of children have cervical spine fractures, and SCIWORA is much more common than in adults.

Spinal Cord Injury Without Radiographic Abnormality (SCIWORA)

This syndrome was first described by Pang and Wilberger in 1982 before the use of MRI to describe SCIs without a discernable fracture on conventional radiographs, CT scans, myelograms, or dynamic flexion–extension radiographs. It excludes injuries caused by penetrating trauma, electric shock, obstetric complications, or those associated with congenital anomalies. In essence, physiologic disruption of the spinal cord occurs without a demonstrable anatomic lesion. The incidence of SCIWORA in spinal cord–injured patients from birth to 17 years is approximately 35%, and the majority occurs in those younger than 8 years. The injury is most common in the cervical region (74%) and may be related to the greater elasticity of the pediatric spinal column, where considerable deformation may occur without disruption.

MRI may be diagnostic in demonstrating spinal cord edema or hemorrhage, soft tissue or ligamentous injury, or apophyseal or disk disruption but may have completely normal results in approximately 35% of cases. SCIWORA is the cause of paralysis in approximately 20% to 30% of all children with injuries of the spinal cord. Potential mechanisms of SCIWORA include hyperextension of the cervical spine, which can cause compression of the spinal cord by the ligamentum flavum, followed by flexion, which can cause longitudinal traction, transient subluxation without gross failure, or unrecognized cartilaginous end plate failure (Salter–Harris type I fracture). Regardless of the specific mechanism, injury to the spinal cord occurs because of the variable elasticity of the elements of the spinal column in children. Experimentally, it has been shown that the osteocartilaginous structures in the spinal column can stretch about 2 inches without disruption but that the spinal cord ruptures after only ¼ inch of elongation. A study of the tensile properties of the pediatric cervical spine noted normalized displacements from the occiput to C2 that were six times greater in an infant than in an adult. The malleable nature of the spinal column is due to the elasticity of the ligaments and joint capsules, high water content of the intervertebral disk and annulus, horizontal facets and wedge-shaped vertebral bodies (permits translation), underdeveloped uncinate processes (allows excessive rotation), and cartilaginous vertebral end plates. In contrast, the spinal cord is relatively tethered by the horizontally departing spinal nerve roots, the dural attachment to the foramen magnum, and the brachial plexus. SCI occurs when deformation of the musculoskeletal structures of the spinal column exceeds the physiologic limits of the spinal cord. Injury may be complete or incomplete. Partial spinal cord syndromes reported in SCIWORA include Brown–Séquard, anterior and central cord syndromes, and mixed patterns of injury.

SCIWORA may also occur in the thoracolumbar spine in association with high-energy thoracic or abdominal trauma. The mechanisms of injury include a vascular insult to the “watershed” area of the spine associated with profound and/or prolonged hypotension, a distraction mechanism in the seat belt–restrained patient, or a hyperextension mechanism after a crush injury as most often occurs when a child is rolled over by a car while in the prone position, resulting in the spine collapsing into the chest cavity.

The prognosis after SCIWORA is correlated with MRI findings, if any are present, and the severity of the neurologic injury. ^∗

∗

Grabb and Pang identified five patterns of intraneural injury that may be seen on MRI. Complete disruption of the spinal cord and cord hemorrhage involving more than 50% of the cross-sectional area of the cord on axial MRI is usually associated with severe neurologic deficits and a dismal prognosis. Minor cord hemorrhage involving less than 50% of the cross-sectional area of the cord is usually associated with moderate to severe neural deficits and a reasonable chance for partial recovery. Edema, without evidence of hemorrhage, is predictive of a very good outcome. No MRI changes are detectable in up to 35% of patients with electrophysiologically proven SCI, and these injuries are associated with an excellent prognosis for full recovery.

Effective management of SCIWORA demands not only careful evaluation of the spine for osseous or cartilaginous injuries or mechanical instability but also stabilization of the spine for prevention of recurrent injury. Radiographic findings of cervical or thoracic fracture–dislocations may be subtle, and errors on the initial reading of radiographs may be as much as 10%. Brace immobilization is usually adequate treatment for SCIWORA. The rationale for brace treatment is that the energy responsible for the injury to the spinal cord also stressed the restraining structures (interspinous ligaments and facet joint capsules) enough to cause a partial tear or severe sprain and occult instability. This concept of occult instability is strongly suggested by two patterns of injury seen in patients with SCIWORA: delayed-onset neurologic deterioration and recurrent SCIWORA. Children with SCIWORA may be seen without neurologic deficit and subsequently develop symptoms up to 4 days after the injury (delayed-onset). Recurrent SCIWORA is characterized by neurologic deterioration that may occur within the first 2 weeks of injury, after an initial period of neurologic stabilization or improvement. Both delayed-onset and recurrent SCIWORA are thought to be caused by repeated injury to a damaged spinal cord that is vulnerable to even minor instability. The incidence of these injury patterns was significantly reduced after initiation of a protocol enforcing strict immobilization of the cervical spine at one institution.

Overall, it is recommended that children with SCIWORA be immobilized for up to 12 weeks and then evaluated with lateral flexion–extension radiographs of the injured region before discontinuing treatment. Cervical and upper thoracic injuries may be treated in a cervicothoracic orthosis (CTO) and lower thoracic and lumbar lesions in a thoracic–lumbar–sacral orthosis (TLSO). Surgical stabilization is rarely indicated. After SCIWORA, a 6-month period of activity restriction is a consideration for prevention of a recurrence. The duration of brace treatment for treatment of SCIWORA remains a controversial issue. A metaanalysis of the literature by Launay and colleagues showed that the chance of recurrent SCIWORA was 17% when brace treatment was discontinued at 8 weeks but that no patients immobilized for 12 weeks developed a recurrent SCI. In contrast, in a 34-year review of SCIWORA at a single institution, recurrent injury was uncommon and of uncertain etiology. Immobilization did not prevent recurrent symptoms or improve outcomes.

Classification of Spinal Cord Injury

Functionally, SCIs are classified as complete if there is total absence of motor and sensory function below the level of injury or incomplete if any neurologic function remains, after resolution of spinal shock. Sacral sparing, evidenced by preservation of sensation in the perianal area (S4–S5 dermatome) and intact anal sphincter control, indicates continuity of long tracts and is associated with a 50% likelihood of improved neurologic function. Spinal cord syndromes rarely are seen as classically described, but being able to define the pattern of neurologic injury may be helpful prognostically. Central cord syndrome is not common in children. It is characterized by more pronounced weakness and sensory changes in the upper extremities than in the lower. Bowel and bladder function is usually unaffected. The chance of recovery is variable and generally better in younger children. Brown-Séquard syndrome is characterized by a functional “hemisection” of the spinal cord with ipsilateral loss of motor and proprioceptive function and contralateral loss of pain and temperature below the level of injury. This injury is more likely to be caused by penetrating trauma, although it has been described in children after blunt trauma. The prognosis is better after blunt injury. Anterior cord syndrome is characterized by loss of motor strength and pain and temperature sensation with preservation of light touch and proprioception. It has the worst prognosis for recovery.

The Frankel classification and American Spinal Injury Association (ASIA) Impairment Scale of traumatic SCIs are both based on testing of motor and sensory function. Frankel grade A is defined by absent motor and sensory function; grade B, by intact sensation but absent motor function; grade C, by intact sensation and motor function active but not useful; grade D, by intact sensation and motor function active but weak; and grade E, by normal motor and sensory function. The ASIA Impairment Scale score, which has gained more widespread acceptance, is based on sensory examination of 28 dermatomes and motor testing of 10 key muscle groups, bilaterally. A sensory score of 0 (absent), 1 (impaired), or 2 (normal) is assigned to each dermatome, and a motor score of 0 to 5 is assigned to each key muscle group. The sensory level is defined as the most caudal level with intact (2/2) sensation and the motor level as the lowest muscle group with at least grade 3 (antigravity) strength. The ASIA Impairment Scale further classifies an injury as complete or incomplete ( Table 12-3 ).

Management of Spinal Cord Injury

After acute SCI, hemodynamic and respiratory monitoring is recommended for detection of cardiovascular or respiratory insufficiency. The rationale is that avoidance of hypoxemia may exacerbate spinal cord ischemia after injury. Clinical series of adult patients treated with aggressive management of blood pressure, oxygenation, and hemodynamics suggest improved neurologic outcomes after SCI without negative effect. At the authors’ institution, patients with acute SCI are managed in the ICU with age-specific mean arterial pressure (MAP) goals over the first 5 days after injury ( Table 12-4 ). Pharmacologic agents are used to maintain the MAP goals, and respiratory insufficiency is avoided with ventilator support, when required.

Pharmacologic treatment of SCI is directed at the secondary injury cascade. The primary insult, usually caused by rapid spinal cord compression in the setting of fracture or dislocation, is irreversible. The hallmarks of the secondary injury cascade include lipid peroxidation, ischemia, and electrolyte derangements. Corticosteroids are stabilizing agents that decrease edema and protect the cell membranes by scavenging oxygen free radicals. Studies in older children (>13 years) and adults indicated that methylprednisolone administered in the first 8 hours after injury may improve the chances of recovery. However, the outcomes of neurologic improvement have not been reproducible in subsequent studies. Additionally, significant complications are associated with high-dose steroids in this population, including hyperglycemia, gastrointestinal bleeding, sepsis, myopathy, urinary tract infections, wound infections, pneumonia, and respiratory failure. ^∗

∗

Enhancement of neurologic recovery after acute SCI has also been reported with administration of GM ₁ ganglioside. This complex acidic glycolipid found in cell membranes in the central nervous system has been shown to have neuroprotective and neurofunctional restoration potential. Prospective, randomized, placebo-controlled drug studies with GM ₁ ganglioside and methylprednisolone have shown improved recovery in patients who were administered both drugs compared with those given methylprednisolone alone. However, a subsequent multicenter randomized study demonstrated no difference in the primary outcome measures at 1 year between the active and placebo treatment groups. Currently, this type of pharmacologic therapy is not recommended.

Enthusiasm for hypothermic therapy has resurfaced after a well-publicized use of systemic hypothermia in a professional football player. Potential benefits of hypothermia include reduced tissue metabolism and energy requirements, slowing of enzymatic activity, decreased parenychmal and axonal swelling, reduced hemorrhaging, and diminished activation of postinjury inflammatory cascades. Levi and colleagues published a retrospective case–control study comparing the complications and outcomes of 14 patients with acute cervical SCI treated with moderate (33°C) intravascular hypothermia with a matched control group of 14 patients. The age range was 16 to 73 years and all had complete SCI at presentation (ASIA grade A). Six patients in the treatment group (43%) improved by one or more ASIA levels in comparison with three in the control group (21%), which was a difference that was not statistically significant. The rate of complications was similar in the two groups, although a statistically greater incidence of pleural effusions and anemia was seen in the treatment group. A potential confounding variable within the study was that early surgery was performed in 85% of the hypothermia group as compared with 50% of the control group. Additionally, postinjury hemodynamic control was not compared between the groups. Systemic hypothermia is a potential therapy for acute SCI that deserves further investigation in well-controlled studies that minimize confounding variables.

The indications for operative treatment in children with SCI are the same as in adults: spinal cord compression in the setting of an incomplete or progressive neurologic deficit, open spinal injury, or a grossly unstable injury pattern. Laminectomy alone is not beneficial. Multiple studies have shown it to be potentially harmful because it increases instability of the spinal column and the likelihood of angular deformity at all levels. The goal of surgical decompression is to prevent or halt the deleterious effects of ischemia, which is an important component of the secondary injury cascade. The standard practice has been to proceed with urgent decompression for incomplete SCIs or those with progressive neurologic deficits. Early decompression (<24 hours after injury) was compared with late decompression in the Surgical Timing in Acute Spinal Cord Injury Study (STASCIS). This was a multicenter, prospective cohort study of patients aged 16 to 80 years with traumatic cervical SCIs (ASIA grade A–D). Improvement of two ASIA grades or greater was noted in 20% of early decompression patients as compared with 9% of the late decompression group ( P = 0.03), but the rate of complications was not statistically different between groups. The STASCIS study suggests improved neurologic outcomes with early decompression of both complete and incomplete SCI. The goal of surgical stabilization is to prevent further mechanical injury to the spinal cord that might exacerbate existing SCI and to permit early mobilization of the patient so that pulmonary complications, disuse osteopenia, and complications from insensate skin are prevented ( Fig 12-12 ). In a study by Jacobs and colleagues comparing operative and nonoperative treatment of spinal fractures in patients with SCI, patients who were treated surgically walked 4.6 weeks earlier or used their wheelchair 5.2 weeks earlier than the nonoperative group.

A 14-year-old male, injured when struck by a car while riding his skateboard. He was seen with an incomplete spinal cord injury with an extension-distraction injury with retrolisthesis of C4 on C5. A and B , Associated injuries included a skull fracture. Given the canal stenosis to 8 mm, he underwent C5 corpectomy and anterior fusion from C4–C6 with a fibular allograft. C , At 2 years postinjury, he had returned to sports with a remaining residual deficit of mild unilateral wrist weakness.

Prognosis and Outcomes of Spinal Cord Injury

Fortunately, children have a greater propensity for neurologic recovery after SCI than adults, and younger children have a better prognosis than older ones. The likelihood for patients with SCI to have at least partial recovery after a complete injury has been reported to be between 10% and 38%. Hadley and colleagues found that 89% of children with incomplete SCIs showed improvement, and 20% of patients with complete SCIs had significant recovery. In Rang’s series, children with both complete and incomplete SCI showed some recovery, except when the injuries occurred in the thoracic region, in which case the prognosis was regarded as hopeless. In the series reported by Wang and colleagues, 64% of their patients demonstrated at least partial recovery after SCI, including 80% with incomplete injuries and 25% with complete injuries who eventually became ambulatory. The early mortality rate was high among those with complete injuries (35%); if only survivors were analyzed, 38% of those with complete injuries had neurologic improvement and progressed to ambulation. Recovery was seen up to 1 year after the original injury and was postulated to be due to the increased ability of the young nervous system to reroute neural pathways and produce axonal sprouting.

Long-term, functional outcomes are determined by the level of SCI. Patients with injuries above C4 usually have paralysis of the diaphragm and often are ventilator dependent. Patients with injuries above C3 can shrug their shoulders and are only capable of neck motion. They can only operate equipment with sip and puff, voice activation, eyebrow or eye blink, or head and chin controls. Patients with C6 lesions can operate the controls of a motorized wheelchair and may be able propel a manual wheelchair if they have triceps function. Children with acquired quadriplegia and paraplegia require management by a variety of specialties including orthopaedics, urology, pediatrics, psychology, social services, physical and occupational therapy, orthotists, and education specialists.

Paralytic, posttraumatic spinal deformity is a complication that is unique to children after an SCI. The combination of spine trauma, paralysis, and growth invariably leads to the occurrence of spinal deformity in a number of ways. The vertebral apophyses in children are functional growth plates that can be injured directly by injury to the spinal column (intrinsic mechanism) or indirectly by factors associated with SCI (extrinsic mechanism). Intrinsic factors such as unstable fractures, loss of ligamentous integrity, and osteochondral (apophyseal) growth plate injuries can contribute to the structural incompetence of the spine, resulting in acute progressive deformity at the fracture site. This mechanism is more typically seen in children who are injured after their teenage growth spurt. Extrinsic factors such as trunk muscle weakness, spasticity, and contractures can cause asymmetric forces to be exerted on the growing vertebral column that result in progressive deformity. This mechanism represents the classic scenario of posttraumatic spinal deformity. A third mechanism by which posttraumatic spinal deformity occurs is iatrogenic and is most often caused by laminectomy without fusion or improperly instrumented spinal levels. The risk of postlaminectomy kyphosis is about 50%, and a much higher incidence occurs in the cervical and thoracic spine.

Patient age at the time of injury and level of paralysis seem to be the most significant factors in the determination of who will develop this problem. Mayfield and colleagues reviewed 49 children younger than 18 years with SCI. Spinal deformities developed in all 28 patients who were injured before their teenage growth spurt had occurred, and the deformities were progressive in 80%. Scoliosis was the most common deformity and was encountered in 93% of patients followed by kyphosis (57%) and lordosis (18%). Approximately 61% of patients required spinal fusion. Conversely, in the group of patients who were injured after the teenage growth spurt, only 38% had a significant, progressive deformity, and only one third of these patients required stabilization. In two other studies by Betz and associates, 98% of children who sustained SCI more than 1 year before skeletal maturity developed spinal deformity, and 67% of them ultimately underwent spinal fusion. However, if the injury occurred less than 1 year before skeletal maturity, the child had only a 20% risk of developing scoliosis and a 5% risk of requiring surgery. In a review of 50 patients, Lancourt and colleagues documented a 100% incidence of scoliosis in children injured before 10 years of age, a 19% incidence in those injured between 10 and 16 years of age, and a 12% incidence in those older than 17 years.

The development of significant spinal deformity can result in pelvic obliquity and difficulty sitting, often requiring the use of the upper extremities for support. Pressure ulcers, pain, and difficulty in proper fit and use of wheelchairs are other significant problems caused by paralytic scoliosis. Orthotic treatment has been shown to be unsuccessful in altering the natural history of these paralytic spinal curves but can be helpful as a temporizing measure for delaying surgery during childhood (<11 years) so that adequate spinal growth can occur. However, the use of an orthosis adversely affects independence level and time requirements for functional activities in this population. When curves are very severe (>40° to 45°) and/or stiff and the patients are older than 10 years, surgical stabilization and fusion are indicated for definitive treatment of the problem. Although these children experience a high complication rate, more than 90% achieve a solid fusion. The goals of surgery are to halt curve progression and obtain curve correction so that the spine and pelvis are balanced; this in turn equalizes sitting skin pressure and restores functional use of the upper extremities. Anterior release may be needed if the deformity is severe and rigid. Long fusions should be performed so that adjacent segment kyphosis is prevented; sacropelvic fixation is necessary if pelvic obliquity exists.

Long-term outcomes are suboptimal after pediatric SCI. Anderson and associates interviewed 161 adults who had sustained SCIs as children and found that 64% lived independently and approximately 50% reported life satisfaction. Compared with the general population, they were less likely to live independently, drive independently, or be married. At longer term follow-up, depression was a common problem among adult patients with pediatric-onset SCI and was associated with poorer outcomes and a lower quality of life. The risk of increased annual mortality was 31% higher for pediatric-onset SCI than for adult-onset SCI. Incomplete injuries with minimal deficits led to 83% normal life expectancy as compared with 50% of normal for those with high cervical injuries without ventilator dependence.