pediatric c-spine clearance journal club · instability, and immobilization with a cervical collar....

!!The!August!2015!Journal!Club!will!be!held!Wednesday!August!26th!in!the!Wong!Kerlee!Conference!Center,!from!8am!to!noon.!Breakfast!will!be!provided.!Discussion!of!articles!will!begin!at!roughly!9am.!The!topic!will!be:!!

Pediatric!CKspine!Clearance!!

!

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!How!do!you!clear!the!CKspine!of!this!patient?!There!are!wellKdescribed!approaches!for!clearance!of!the!adult!CKspine!in!the!medical!literature.!Do!they!work!in!children?!!!!

Emergency!Medicine!Journal!Club!Wednesday,!August!26,!2015!

The!following!individuals!have!been!assigned!to!present!articles!(maximum!of!5K10!minutes!each).!Everyone!is!expected,!however,!to!have!read!the!articles!and!to!be!prepared!to!critically!discuss!them.!!!Page! Presenter! Article!! Drew!Davis!and!Nick!

Selden!Pang%D,!Wilberger!JE!Jr.!Spinal!cord!injury!without!radiographic!abnormalities!in!children.!J!Neurosurg.!1982!Jul;57(1):114K29.!!

! Brad!Alice! Kenter%K,!Worley!G,!Griffin!T,!Fitch!RD.!Pediatric!traumatic!atlantoKoccipital!dislocation:!five!cases!and!a!review.!J!Pediatr!Orthop.!2001!SepKOct;21(5):585K9.!

! Darren!Brockie!and!Nathan!Silvestri!

Brown%RL,!Brunn!MA,!Garcia!VF.!Cervical!spine!injuries!in!children:!a!review!of!103!patients!treated!consecutively!at!a!level!1!pediatric!trauma!center.!J!Pediatr!Surg.!2001!Aug;36(8):1107K14.!

! Tim!Young! Stiell%IG,!Wells!GA,!Vandemheen!KL,!et!al.!The!Canadian!CKspine!rule!for!radiography!in!alert!and!stable!trauma!patients.!JAMA.!2001!Oct!!17;286(15):1841K8.!

! Liz!Fierro!and!Jennifer!Hughes!

Viccellio%P,!Simon!H,!Pressman!BD,!Shah!MN,!Mower!WR,!Hoffman!JR;!NEXUS!Group.!A!prospective!multicenter!study!of!cervical!spine!injury!in!children.!Pediatrics.!2001!Aug;108(2):E20.!!

! Tim!Young! Garton%HJ,%Hammer!MR.!Detection!of!pediatric!cervical!spine!injury.!Neurosurgery.!2008!Mar;62(3):700K8.%

! Tim!Young! Stiell%IG,%Wells!GA.!Methodologic!standards!for!the!development!of!clinical!decision!rules!in!emergency!medicine.!Ann!Emerg!Med.!1999!Apr;33(4):437K47.!

! Tim!Young! Burns%EC,!Yanchar!NL.!Using!cervical!spine!clearance!guidelines!in!a!pediatric!population:!a!survey!of!physician!practices!and!opinions.!CJEM.!2011!Jan;13(1):1K6.!!

! !Tim!Young! Pang%D.!Spinal!cord!injury!without!radiographic!abnormality!in!children,!2!decades!later.!Neurosurgery.!2004!Dec;55(6):1325K42;!discussion!1342K3.!

! Mike!O’Neal!and!Harut!Hovsepyan!

Yucesoy%K,!Yuksel!KZ.!SCIWORA!in!MRI!era.!Clin!Neurol!Neurosurg.!2008!May;110(5):429K33.!doi:!10.1016/j.clineuro.2008.02.004.!Epub!2008!Mar!19.!

! Erik!Smith!and!Sarah!Peterson!!

Leonard%JC,%Kuppermann!N,!Olsen!C,!et!al.!Pediatric!Emergency!Care!Applied!Research!Network.!Factors!associated!with!cervical!spine!injury!in!children!after!blunt!trauma.!Ann!Emerg!Med.!2011!Aug;58(2):145K55.!

! Allen!Chiou!and!Jason!Morris!

Hale%DF,!Fitzpatrick!CM,!Doski!JJ,!Stewart!RM,!Mueller!DL.!Absence!of!clinical!findings!reliably!excludes!unstable!cervical!spine!injuries!in!children!5!years!or!younger.!J!Trauma!Acute!Care!Surg.!2015!May;78(5):943K8.!

!

3

19

24

32

34

40

41

43

44

45

50

61

J Neurosurg57:l14-129, 1982

Spinal cord injury without radiographic abnormalities in children

DACHLING PANG, M.D., F.R.C.S.(C), AND JAMES E. WILBERGER, JR., M.D.

Department of Neurosurgery, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania

~/ Spinal cord injury in children often occurs without evidence of fracture or dislocation. The mechanisms of neural damage in this syndrome of spinal cord injury without radiographic abnormality (SCIWORA) include flexion, hyperextension, longitudinal distraction, and ischemia. Inherent elasticity of the vertebral column in infants and young children, among other age-related anatomical peculiarities, render the pediatric spine exceedingly vulnerable to deforming forces. The neurological lesions encountered in this syndrome include a high incidence of complete and severe partial cord lesions. Children younger than 8 years old sustain more serious neurological damage and suffer a larger number of upper cervical cord lesions than children aged over 8 years. Of the children with SCIWORA, 52% have delayed onset of paralysis up to 4 days after injury, and most of these children recall transient paresthesia, numbness, or subjective paralysis. Management includes tomography and flexion-extension films to rule out incipient instability, and immobilization with a cervical collar. Delayed dynamic films are essential to exclude late instability, which, if present, should be managed with Halo fixation or surgical fusion. The long-term prognosis in cases of SCIWORA is grim. Most children with complete and severe lesions do not recover; only those with initially mild neural injuries make satisfactory neurological recovery.

KEY WORDS �9 spinal cord injury juvenile spine increased deformability severe neurological damage �9 delayed paralysis �9 tomography �9 dynamic radiography �9 children

B ECAUSE of anatomical and biomechanical differences in the human spine at various ages, the mechanism of neural damage and degree

of osteoligamentous disruptions associated with spinal cord injury may be radically different from one age group to another. Accordingly, patients with spinal cord injury can be divided into four age categories: those injured 1) during birth; 2) between infancy and 16 years; 3) between 16 years and middle age; and 4) between late-middle to old age.

The mechanism involved in intrapartum spinal cord injury is generally thought to involve longitudinal traction of the neonatal spine during breech extraction and subsequent rupture of the cord. 126,48 Due to the extreme elasticity of the fibrocartilaginous spine and its investing soft tissues, the resulting tetraplegia is often described without radiographical evidence of fracture or dislocation. 1,2~,~1

From 16 years to middle age (arbitrarily defined as 45 years), it is exceedingly rare to find closed spinal cord trauma without skeletal injury or dislocation. 52,62

The mechanical properties of the spines of patients in this age range are such that fractures of the vertebral body, pedicles, or facets, or locking of an anteriorly dislocated facet occur before the neural structures are injured.

In patients of late-middle to old age, a syndrome of closed spinal cord injury without demonstrable skeletal injury was identified by Cr0oks and Birkett in 1944. 33 The spines of these patients usually show significant spondylotic changes, resulting in sagittal narrowing of the central canal, and the neurological injury is most often acute central cord syndrome. The mechanism of injury has been convincingly shown by Taylor and Blackwood 71,72 and Schneider, et aL, 66'67 to be hyperextension.

Spinal cord injury is uncommon from infancy to 16 years. The incidence of pediatric spinal cord injuries among all spinal cord injuries has been quoted as anywhere from 0.65% to 9.47%) 6,39,~~ Although many children with spinal cord injury have associated vertebral injuries, a distinct group of children with

1 1 4 J. Neurosurg. / Volume 57 / July, 1982

3

Cord injury with normal radiography

traumatic myelopathy have no radiographic evidence of fracture or dislocation, z,23,24,31,42,44,46,5~176

Twenty-four children with this syndrome of spinal cord injury without radiographic abnormality (SCIWORA) were treated at Children's Hospital of Pittsburgh (CHP) from 1960 to 1980, constituting 66.7% of all children with nonpenetrating spinal cord injuries seen at this institution during that time. Our experience shows that the pediatric syndrome of SCIWORA is characterized by clinical features and prognosis very different from those found in children with spinal cord injury and associated fracture-dislocation, or in adults with cervical spondylosis and hyperextension injury. It is probable that complex pathogenetic mechanisms unique to the spine of young children are involved in producing this syndrome. It is this specific subgroup of pediatric spinal cord trauma that this paper wishes to address.

Clinical Material and Type of Injury Definition of the Series

The pediatric syndrome of SCIWORA is reserved for those children with objective signs of myelopathy as a result of trauma, whose plain films of the spine, tomography, and occasionally myelography carried out at the time of admission showed no evidence of skeletal injury or subluxation. We have excluded all birth injuries and children with cord injuries caused by electric shock, penetrating agents, or missiles. Since we were primarily interested in the vulnerability of the normal spine in the young to violence, we have also eliminated all congenital malformations associated with inherent instability of the spine, such as insufficiency of the transverse odontoid ligament, os odontoideum, ossiculum terminale, the Klippel-Feil syndrome, Down's syndrome, and occipitalization of the atlas.

Age and Sex of Patients From 1960 to 1980, 36 children with closed spinal

cord injuries were treated at CHP. Twelve children had radiographic evidence of fracture or fracture- subluxation, and 24 belonged to the SCIWORA group. Thus, the SCIWORA group constituted 66.7% of the total.

The age range varied from 6 months to 16 years, with a mean of 7.2 years. Fourteen children were younger than 8 years (58.3%), and 10 were aged from 8 to 16 years (41.7%). There were 10 males and 14 females. The follow-up period ranged from 8 months to 20 years.

Cause of Injury

The most common causes of injury were related to vehicular accidents (Table 1). Four children were in the front seat and one in the back seat during colli- sions. One 7-year-old child was thrown from the back seat of a motor cycle. The four children hit by auto-

TABLE 1 Cause of injury in 24 children

Cause of Injury Cases

No. Percent

hit by car 4 16.7 run over by car (chest) 1 4.2 automobile accident 5 20.8 motorcycle accident 1 4.2 fall from height 5 20.8 fall down steps 2 8.3 football tackle 1 4.2 diving 1 4.2 object fell on head 1 4.2 sled accident 1 4.2 wrestling 1 4.2 child abuse 1 4.2

mobiles were all under 6 years of age. A 16-month- old child was run over on the chest by his father's truck as it was being backed out from the garage.

The next most common causes involved falls from height: one from a crib, one in the gymnasium, and three from trees. Most of these children were under 8 years old. Two toddlers fell down steps at home by escaping the confines of their walkers. The other isolated causes were mainly sports-related, and all involved older children. There was one case of child abuse.

Mechanism of Injury

In most cases, the mechanism of neural injury could be deduced from associated bone and soft-tissue injuries (Table 2). Chin laceration, mandibular fracture, facial injuries, and frontal fracture or bruising usually implied hyperextension; conversely, an occipital bruise, laceration, or fracture pointed to a flexion-type injury. These clues were particularly important in infants and young children, and in those children whose concomitant concussion or traumatic amnesia had rendered a coherent history impossible. In some cases involving older children, surface clues were corroborated by good descriptions from the patient or an eye witness. A hyperextension sprain was almost certainly involved when an 11-year-old child suffered a cervical cord injury while doing backward somer- saults. A flexion mechanism was suspected in a 9- year-old boy who described the onset of symptoms occurring when his neck was forced into extreme flexion by an occipital blow during a wrestling match.

A 1�89 victim of child abuse had subhyaloid hemorrhages, subarachnoid hemorrhage, and diffuse cerebral swelling in addition to a partial C-5 myelopathy. The known association of a whiplash-shake type of violence with this form of cerebral lesion suggested that a combination of repetitive hyperextension and flexion forces caused this child's spinal cord damage.

Two children had flexion-compression injuries to their cervical spine; one as a result of diving into

J. Neurosurg. / Volume 57 / July, 1982 11 5

4

TABLE 2 Mechanism of injury and diagnostic clues in 24 children

Age Mechanism Diagnostic Clues (yrs) 2�89 hyperextension depressed frontal fracture 3" hyperextension chin laceration; mandibular

fracture 3 hyperextension frontal laceration 8�89 hyperextension forehead bruise

10 hyperextension patient's description 11 hyperextension backward somersaulting 14 hyperextension patient's description (football

tackle) 15 hyperextension forehead & facial lacerations 15" hyperextension mandibular fracture; anterior

neck lacerations; chin abrasion 16 hyperextension rt frontal, facial lacerations

�89 flexion mother's description (fell down steps)

2 flexion occipital bruise 2�89 flexion occipital bruise 4 flexion occipital bruise 4 flexion occipital bruise 6* flexion occipital bruise 7" flexion occipital laceration & fracture 9 flexion struck on occiput while wrestling 1�89 repetitive flexion child abuse: shake & whiplash

& extension injury 8 flexion-compres- hit vertex while diving

sion 10 flexion-compres- heavy object fell on vertex

sion 4 longitudinal head caught by tow chain,

distraction dragged 5* longitudinal lap seat belt injury; associated

distraction L2-3 transverse fracture 13 direct crush run over by truck, lying on

injury ? hyper- abdomen; tire marks extension

* Severe hypotension (systolic pressure < 60 torr) present at admission.

shallow water, and one from a direct vertex hit from a falling object. Both had signs of soft-tissue bruising in the vertex.

A 5-year-old child was strapped to the front passenger seat by a lap seat belt during a head-on collision. He suffered a transverse body fracture of the L-2 vertebral body (Chance fracture) and a complete mid- thoracic transverse myelopathy. Since a flexion-distraction force of great magnitude is necessary to produce the type of lumbar fracture in question, the same force vector may account for his thoracic cord injury. Similar distraction forces were likely responsible for the cervical cord injury of a 4-year-old child whose head was caught by a tow chain which dragged his body a considerable distance.

There was one case of direct crush injury to the thoracic spine in a 16-month-old child who was run over by a truck while lying on his abdomen, evidenced by fresh tire markings on his back. The forces involved would tend to cause a backward bending on the spine.

Flexion is as frequently implicated as hyperextension as the mechanism of injury (Table 2). However,

D. Pang and J. E. Wilberger, Jr.

if the mechanism of injury is correlated with age, flexion and hyperextension appear to involve two different age groups. Of the eight children who sustained flexion injury, seven were younger than 8 years. Conversely, seven of the 10 with hyperextension injury were older children. Both cases of longitudinal distraction and the one case of direct crush injury to the thoracic spine occurred in very young children.

Table 2 also indicates six instances where severe hypotension (systolic pressure less than 60 torr) was present at the time of admission. In all six cases, this was due to blood loss from thoracoabdominal injuries, multiple long-bone fractures, or soft-tissue lacerations.

Associated Extraneural Injuries

Three children had intra-abdominal hemorrhage caused, respectively, by a mesenteric tear, a ruptured spleen, and a liver laceration; all three had significant hypotension at the time of admission. One child was in hypovolemic shock because of a severe oropharyn- geal tear and another because of a massive avulsion of the occipital scalp over a depressed fracture.

Four patients had skull fractures: two frontal, one occipital, and one basilar. Both children with frontal fractures also had associated mandibular fractures. Two children had femoral-tibial fractures and one had pelvic fractures.

Neurological Status

Level and Type of Neurological Lesions

Five lesions involved the upper cervical cord (C1-4), 15 involved the lower cervical cord (C-5 and C-8), and four involved the thoracic cord.

Neurological examination at the time of admission revealed four individual syndromes: complete physiological cord transection, central cord syndrome, Brown-S&luard syndrome, and partial cord syndrome (Table 3). The latter category included those patients with partial preservation of function below the level of the lesion, but whose pattern of neurological deficits could not be classified as either central cord or anterior cord syndromes.

There were seven cases of complete transection. Four of the ten patients with central cord syndrome were classified as severe because of profound weakness of hand grip and forearm musculature and sufficient weakness of the lower extremities to prevent ambulation. The other six children had a mild central cord syndrome, and either remained ambulatory or showed rapid improvement in lower extremity func- tions within the first 24 hours. The patient with Brown-S&luard syndrome had a hemihypalgesia and mild weakness in the contralateral arm, and a corresponding Babinski response. There were three patients each with severe and mild partial cord syndrome. In all, 14 children (58.3%) either sustained complete physiological transection or serious damage to their spinal cord.

1 1 6 J. Neurosurg. / Volume 57 / July, 1982

5


TABLE 3 Types of neurological syndromes in 24 children

Neurological No. of Percent of Syndromes Cases Total

complete cord transection 7 29.1 central cord, severe 4 16.7 central cord, mild 6 25.0 partial cord, severe 3 12.5 partial cord, mild 3 12.5 Brown-S6quard 1 4.2

TABLE 4 Correlation between level and severity of neurological injury

Level of No. of Severity of Injury Injury Cases Complete Severe Mild

c1-4 5 3 2 0 c5-8 15 1 4 lO T1-6 4 3 1 0

Level Related to Severity of Neurological Injury

The upper cervical cord and thoracic cord were equally prone to serious injuries (Table 4). Six of the seven complete transections in the series occurred in these two regions. On the other hand, the lower cervical spine appeared to be relatively resistant to injury. Only one case of complete transection and four of the severe cord syndromes involved the C5-8 segments, and 10 of the 15 injuries to this region were mild or moderate.

Age Related to Neurological Status

Figure 1 suggests that the spines in infants and young children were more vulnerable to deforming forces than the spines of older children. Children aged from 6 months to 8 years suffered much more devastating neurological injuries: seven of the 14 younger children had complete transections, and six had either a severe central cord syndrome or a severe partial cord syndrome. Only one child in this age group who suffered a flexion injury to the cervical spine es- caped serious cord damage. Conversely, all but one of the children aged 8 to 16 years had mild to moderate neurological damage. The one exception was an 11-year-old boy with a nearly complete paraplegia at T-5 following a violent extension injury.

The same age-related difference exists within the subgroup of the 10 cases of central cord syndrome. The four severe cases were 4 years old or younger, and the six mild cases were older than 8 years.

Figure 2 shows the age distribution in relation to the level of neurological injury. All five cases of upper cervical cord damage occurred in the younger group, whereas injuries to the lower cervical segments were evenly distributed in the entire age span of the series.

COMPLETE CORD

SEVERE CENTRAL

SEVERE PARTIAL

MILD CENTRAL

MILD PARTIAL

BROWN SEQUARD

�9 I l l I l l I I

A A i l i

:A AA A A A

o o

I I I I I I I I I n I I n I I n n

8 YEARS 16 YEARS

FIG. 1. Correlation between patients' age and neurological syndromes.

C 1 - C 4

C 5 - C 8

T 1 - T 6

0 0 0 �9 �9

�9 0 0 � 9 �9

I I I I I I L

o o o �9 0 0 �9

8 1 6 YEARS YEARS

FIG. 2. Correlation between patients' age and level of neurological injury.

The deduction can be made that the upper cervical spine is inherently more susceptible to injury in infancy and early childhood but gains stability with increasing age. The same statement may also be applicable to the thoracic spine, although the small number of thoracic cord injuries in this series pre- cludes a definite conclusion. The lower cervical spine, on the other hand, appears to be at risk in all ages within the age limits of this series.

Mechanism of Injury Related to Severity and Level o f Neurological Lesion

Table 5 attempts to define the relationship between the types of deforming forces to the spine and the severity of the resultant neurological damage. Flexion forces to the spine appear to produce more serious neural injuries than extension forces. One explanation is that the pediatric spine is more resistant to extension forces than to flexion forces. However, if the ages of these patients were taken into consideration, it is clear that the three cases of complete cord transection due to hyperextension and the six cases of severe to complete cord injury due to flexion all occurred in children under 6 years. Thus, an alternative explanation may again be related to the aforementioned suggestion that the spines in infants and young children are inherently


6

TABLE 5 Correlation between mechanism of injury and severity of

neurological injury

Mechanism Severity of Injury of Injury Complete Severe Mild

hyperextension 3 1 6 flexion 2 4 2 flexion-extension 0 1 0 longitudinal distraction 1 1 0 direct crush injury 1 0 0 flexion-compression 0 0 2


TABLE 6 Correlation between mechanism of injury and level of

neurological lesions

Mechanism Level of Injury of Injury C1--4 C5-8 T1-6

hyperextension 8 flexion 5 3 flexion-extension 1 longitudinal distraction 1 direct crush injury flexion-compression 2

more deformable and, therefore, provide less effective proteetion for the subjacent spinal cord regardless of whether flexion or extension forces were involved. Certainly, within the subgroup of patients subjected to hyperextension, the younger patients all had worse neural damage than the older ones (Fig. 1).

Repetitive flexion-extension as seen in the whiplash-shake type of child abuse, longitudinal distraction of the spine, and direct crush injury were all associated with severe neural damage (Table 5). How- ever, it must be remembered that these injuries all involved extremely violent forces and were also in- flicted on infants and very young children. Both patients who suffered vertical loading to their vertices had relatively mild neural lesions.

Table 6 relates the mechanism of injury to the level of the neurological lesion. It appears that the upper and lower halves of the cervical cord are susceptible to different types of deforming forces. All five cases of upper cervical cord injuries in the series were caused by flexion forces, while the predominant mechanism injuring the lower four cervical segments was hyperextension. Again, if one includes the age factor in the analysis, it is apparent that if flexion forces are applied to the very young spine, the upper four segments of the cord are most likely to be injured; but in children older than 8 years, both flexion and extension forces are more likely to injure the lower cord segments.

Table 6 also shows that thoracic cord injury can be caused by several different mechanisms. It is probable that what determines the severity of thoracic cord injury is the magnitude and not the direction of the deforming vector.

Delayed Onset o f Neurological Signs

Thirteen patients in this series (54%) had delayed onset of their neurological deficits following spinal trauma (Table 7). The time interval between injury and the appearance of objective sensorimotor dys- function, the "latent period," ranged from 30 minutes to 4 days, with a mean of 1.2 days. There was no uniformity in either the age or mechanism of injury in this subgroup. The age range varied from 6 months to 15 years, and the mechanisms included six cases of

hyperextension, three of flexion, two of flexion-compression, one of longitudinal distraction, and one of direct crush injury to the thoracic spine.

Immediately following injury, the child in Case 1 was noted to have transient clumsiness in his extremities which disappeared within minutes; he became profoundly quadriparetic 2 days later. Seven other children recalled transient neurological symptoms that were initially ignored by the patients and often by their physicians. Four children (Cases 2, 6, 7, and 8) had paresthesia in all extremities lasting from 5 minutes to 1 hour. A 10-year-old girl who suffered a flexion-compression injury from diving into shallow water (Case 5) had transient tingling in both upper extremities and had a subjective feeling of"total body paralysis," although no objective deficits were found. She developed a severe central cord syndrome 4 hours later. In Case 3, the child described a similar feeling of subjective paralysis, was examined by a neurosur- geon, found to be normal, and sent home. He was well for 2 days but, while playing basketball 48 hours following his injury, he rapidly developed arm and hand weakness and paresthesia in both legs. The child in Case 4 recalled a tingling numbness in his hands and a lightning sensation shooting down both legs, together with a subjective "heaviness" in the lower part of his body. These symptoms cleared within 15 minutes, but 24 hours later he presented with a typical C-5 central cord syndrome. During his 2-day hospital stay, his weakness improved dramatically, and he was sent home with a stiff cervical collar. Two days later, he removed his collar while playing and promptly noticed symptoms caused by a recurrence of his myelopathy.

Although these patients were well during the "latent period," sensorimotor paralysis progressed inexorably once it began, evolving into complete transverse myelopathy in two cases, severe partial cord syndrome in four, and mild to moderate central cord syndrome in seven children.

Radiographic Diagnosis The radiographic tests performed are summarized

in Table 8. All patients had plain films of the spine taken on admission. The entire spine was studied in

118 J. Neurosurg. / Volume 57 / July, 1982

7


Case Age Mechanism of No. (yrs) Injury

1 2 flexion

2 4 longitudinal distraction

3 8�89 extension

4 10 flexion-compression

5 10 flexion-compression

6 11 extension

7 14 extension 8 15 extension

9 �89 flexion

10 1-~ direct crush injury

11 2�89 extension

12 9 flexion

13 16 extension

TABLE 7 Delayed neurological signs in 13 children

Initial Transient Symptoms

mother noted child not moving limbs immediately after injury; cleared quickly

paresthesia, both arms & legs

subjective feeling of paralysis

lightning sensation; paresthesia, both hands; subjective feeling of paralysis

paresthesia both hands, subjective weakness

paresthesia, both legs

"Latent Neurological Myelog- Period" Manifestations raphy Final Outcome

2 days severe C-6 central normal severe deficits cord

2 days severe C-5 central severe deficits cord

2 days mild C-5 central normal cord

24 hrs mild C-5 central normal cord

4 hrs mild C-5 central mild deficits cord

12 hrs severe T-5 partial normal severe deficits cord

12 hrs mild central cord normal normal 6 hrs mild C-6 central normal

cord 4 days severe C-3 partial normal moderate

cord deficits 24 hrs T-6 complete cord normal complete cord

transection transection 4 days C-7 complete cord normal complete cord

transection transection 30 min mild C-5 central mild deficits

cord 24 hrs mild C-5 central normal

cord

paresthesia, both hands & legs paresthesia, both hands

those patients with multisystem trauma, but only the symptomatic area was studied in those with isolated neurological syndromes. All but one patient had normal studies. This child sustained a lap seat belt injury and had a transverse fracture of the L-2 vertebral body (Chance fracture). However, her level o f neural injury was at T-6 and her thoracic spine film was normal.

Tomographies obtained in 83.3% of the patients were normal. Myelography with either Pantopaque or gas was performed on 50% of the patients to rule out a subarachnoid block, usually in the wake of normal plain films and tomography. Free flow of contrast medium past the region of neurological lesion was demonstrated in each case, al though some cords looked slightly swollen with irregular contour.

Computerized tomography (CT) of the spine was performed on one patient who suffered a mild central cord syndrome following a hyperextension injury. Transverse imaging through the C-5 level disclosed a small hyperdense area compatible with an epidural hematoma. No cord compression or bone abnormali ty was appreciated. Repeat CT 2 weeks later was completely normal.

Dynamic films with the patient 's neck in voluntary flexion and extension were obtained in 75% of the children after plain films in the neutral position were found to be normal. No acute instability was demonstrated, but in most cases, paraspinal muscle spasm severely restricted the range of motion and rendered these studies technically inadequate. Delayed dy-

TABLE 8 Radiographic tests performed in 24 patients

Radiography Cases No. Percent

plain spine films 24 100 tomography 20 83.3 myelography 12 50 dynamic studies

acute 18 75 delayed 24 100

computerized tomography 1 4.2 of spine

namic studies were performed on all patients several days after admission, at a time when the neurological status had stabilized or after muscle spasm had subsided. Instability was demonstrated in only one patient who showed mild anterior slipping of C-4 and C-5. This patient had severe spasm during the 1st week of admission and had inadequate acute flexion- extension studies.

Treatment and Outcome

Management

All patients with cervical cord syndromes had immediate neck immobilization with cervical collars. Those with thoracic cord injuries were placed supine on a fracture board before other resuscitative measures were instituted.

J. Neurosurg. / Volume 57 / July, 1982 1 1 9

8

The six patients in hypovolemic shock were treated with vigorous fluid and blood replacement to avoid persistent systemic hypotension. Emergency tracheos- tomies were performed on the five patients with upper cervical cord injuries and mechanical ventilation was begun. Corticosteroids were routinely used on all patients. In-dwelling bladder catheters were inserted into those patients with complete or severe partial cord syndromes during the acute phase of injury, and intermittent catheterization programs were set up for those infants and young children who required long- term assistance in bladder drainage.

Intracranial pressure monitors were used on two children with severe concomitant head injuries. They were subsequently managed with our intracranial hy- pertension protocol in the intensive care unit (ICU). Three children had laparotomies for intra-abdominal hemorrhage, and two other children had repair of the oropharynx and an occipital scalp avulsion, respectively. All long-bone fractures were managed initially with skeletal traction.

Following the radiographic establishment of the SCIWORA syndrome, all thoracic cord injuries were treated with bed rest for 1 week; stability was confirmed by follow-up films of the thoracic spine, and the patients were mobilized. One patient with a transverse lumbar fracture had posterior Harrington rod fusion. Rehabilitative measures began immediately following admission to the hospital.

Patients with high cervical cord lesions required prolonged mechanical ventilation and ICU care. They were kept in cervical collars for 1 month, had biweekly cervical spine films with delayed flexion-extension views to rule out subtle instability, and received early rehabilitative intervention. Those with severe central and partial lesions of the lower cervical cord were similarly immobilized with a cervical collar while on the neurological unit, and those without late instability were transferred to the rehabilitation unit after 1 week of bed rest. The child with the C-4 subluxation found on delayed flexion-extension studies was immobilized with the halo apparatus for 8 weeks and subsequently achieved stability. Patients with mild injuries who remained ambulatory or showed substantial clinical improvement within the first 48 hours were allowed ward privileges in the 1st week and discharged home on the following week wearing their cervical collar. This was kept on for 4 weeks, at which time a second set of dynamic films was always obtained to ensure stability.

Complications

The two most common complications during the acute treatment period were frequent respiratory and urinary tract infections. There was one case of deep vein thrombosis in a child with a complete C-6 transection.

Late complications included distressing involuntary muscle spasm below the level of the lesion and serious

D. P a n g and J. E. Wi lbe rge r , Jr.

psychological maladjustment in the severely disabled children. No cases of delayed spinal deformity were encountered in this series.

Outcome

For children with SCIWORA, the long-term prognosis is poor (Table 9). One child with a complete cord syndrome died. This 6-year-old boy was hit by a car and sustained massive thoracoabdominal injuries. He presented with profound hypovolemic shock, a complete C-2 cord transection, and cardiorespira- tory arrest. Despite heroic efforts at resuscitation, he died of progressive respiratory failure 4 days later.

The six remaining patients with complete cord syndromes remained unchanged neurologically. Two of these children with upper cervical cord injuries required long-term mechanical ventilation. One child with a C-6 complete cord syndrome spent 2 years in a rehabilitation institute but could not attain the level of self-care, partly because of troublesome muscle spasms and frequent respiratory complications.

The three children with initially complete thoracic cord transection made no neurological recovery but proved to be far better rehabilitation candidates than patients with severe cervical injuries. Five of the seven children with severe cord syndromes continue to have severe deficits. Since most of these had cervical cord injuries, they represent a severely handicapped group. Only two children in this group made a substantial recovery to the point of being ambulatory with pros- theses.

The 10 children with initially mild to moderate neural damage represent the only optimistic group in the series. Seven were neurologically normal 3 months to 7 years following initial hospitalization. The three others have residual deficits but are enjoying a full psychosocial lifestyle.

It appears from these figures that the most important factor determining prognosis is the initial neurological status. Twelve of 14 children (85.7%) with initially complete or severe cord syndromes either are dead or continue to be totally disabled. Only two children (14.2%) from this group made satisfactory progress. On the other hand, seven of the 10 patients (70%) with mild to moderate initial damage to the cord have made complete recovery, and the rest are only minimally disabled.

Discussion

Because of the rarity of pediatric spinal cord injuries, comprehensive clinical and epidemiological data concerning the syndrome of spinal cord injury without radiographic abnormality (SCIWORA) are not found in the literature. However, the phenomenon of SCIWORA in children has been noted by several investigators. In 1969, Audic and Maury 6 stated that in 21 children under the age of 16 years with spinal

1 20 J. Neurosurg. / Volume 57 / July, 1982

9


TABLE 9 Outcome in 24 children

Initial Neurological No. of Status Cases Death

Final Neurological Status

Complete Severe Mild/ Cord Syn- Deficits Moderate

drome Deficits Normal

complete cord syndrome 7 severe cord syndrome 7 mild cord syndrome 10

cord trauma, "very often" no fracture was detected. In the same year, Melzak 6~ reported 16 similar cases among 29 children with spinal cord injury. The num- bers of children with SCIWORA in other series of pediatric spinal cord trauma are as follows: Burke 24 found 12 out of 24; Hasue, et al., 46 one out of 10; Hachen 44 eight out of 18; Andrews and Jung 5 seven out of 15; Anderson and Schutt 4 two out of 42; and Kewalramani, et al., ~~ five out of 25 children. Unfor- tunately, these data were obtained from records of rehabilitation facilities and lack detailed clinical information concerning the individual patients.

Some insight into this problem was provided by four additional articles which documented the clinical courses of 13 children with traumatic myelopathy without fracture or dislocation. 2,23,31m Our own figure of 24 cases of SCIWORA among 36 children with spinal cord injury convinces us that this is a common entity in pediatric spinal trauma. Based on others' experience and the information gathered from the present study, aided by the current knowledge regarding the anatomical preculiarities of the pediatric spine and the biomechanics of spinal cord injury in general, we examined the mechanism of injury, the mode of presentation, the neurological lesions, the management, and the prognosis of this syndrome.

Mechanisms o f Neural Damage Without Fracture or Dislocation

Most of the existing knowledge on the biomechanics of spinal injury is based on studies of the adult spine. Some of this work has been inspired by a need to explain the well known phenomenon of acute central cord syndrome in the elderly without evidence of vertebral fracture or dislocation. A search for the mechanism of the childhood syndrome of SCIWORA must, therefore, borrow heavily from the results of the adult studies, taking into due consideration the anatomy and kinetics of the juvenile spine. It is probable that more than one mechanism is involved in the pathogenesis of SCIWORA.

Hyperextension Injuries

Based on cadaver study, Taylor in 195171 demonstrated that the interlaminar ligaments of the cervi-

cal spine bulged forward into the central canal during hyperextension. Hyperextension was therefore thought to be the cause of the acute central cord syndrome often seen in the elderly without evidence of fracture. This theory was supported by Alexander, et al., 3 who found that the sagittal diameter of the cervical canal could be narrowed by over 50% during extension. They claimed that if preexisting spondylotic protrusions were present to further narrow the canal at C4-6 to less than 13 mm during hyperextension, the spinal cord would be pinched between the osteophytes and the inward bulging interlaminar ligaments. An additional aggravating factor was later found by Breig and E1-Nadi 22 to be a shortening and, therefore, thickening of the cord during hyperextension as the spinal column shortens. This would accen- tuate the crowding of the intraspinal contents.

In one case of hyperextension injury involving extreme violence, Taylor and Blackwood 72 found at necropsy that the anterior longitudinal ligament had ruptured, the intervertebral disc had detached from the lower vertebral body, and the segment of the cervical column above this disc had become displaced backward to compress the cord. However, elastic recoil of the paraspinous muscles resulted in spontaneous reduction of the displacement TM and gave a normal radiographic appearance. They postulated that with hyperextension sprain exceeding the tensile resistance of the anterior longitudinal ligament, this structure, unattached to and hence unsupported by the anulus, 2~ will rupture and permit retrolisthesis of the upper segments. Anatomical plausibility of this mechanism was provided by Bourmer, 17 who showed on five cadavers that if the anterior longitudinal ligament was severed and if a backward displacing force was directed at the head, the intervertebral disc promptly ruptured to allow for retrolisthesis of the body above. Radiographic confirmation of the Taylor-Blackwood mechanism came in 1974 when Marar ~6 showed subtle retrolisthesis in 11 of 45 patients suffering from quadriplegia due to hyperextension. Using postmortem stress radiography on two other patients who succumbed to this injury, he demonstrated opening of the disc space and backward displacement of the upper body. These tissue injuries were later confirmed at autopsy.


10

Thus, it may be assumed that with a moderate degree of hyperextension, the cervical cord can be compressed by the inward bulging of the interlaminar ligaments against preexisting spondylosis or congenital stenosis, aggravated by a thickening of its cross section area during its simultaneous shortening. Ex- ceptionally violent hyperextension will rupture the anterior longitudinal ligament, detach the disc, and cause a retrodisplacement of the upper body. Imme- diate muscle action causes elastic recoil and spontaneous reduction, and reflex muscle spasm maintains the relatively stable reduction to give the normal radiographic appearance. 16'1s'37'61'62

Several anatomical characteristics of the pediatric spine increase its susceptibility to hyperextension injury. In children, the ligaments, posterior joint capsules, and cartilaginous structures are more elastic than the rigid adult spine and, therefore, more deformable. 9'3~ The planes of the facet joints in children are also more horizontal and thus possess greater mobility but less stability. 9,3~ In addition, Aufdermaur 7 found that in the young spine, a locus of weakness exists within the growth zone of the cartilaginous end plate at its junction with the primary centrum, where the fibrous lamellae are loosely ar- ranged, being segregated by columns of cartilage cells. Separation of the end plate from the centrum readily occurs at this region with even a moderate degree of shearing, as evidenced in 12 autopsy studies. Two patients described by Cheshire 31 and one by Ahmann, et al.,2 probably had hyperextension. Hyperextension is implicated in 10 children in the present series. Scher 64 showed that hyperextension was maximum at the C5-6 junction, and suggested that hyperextension injury to the cord most likely occurred in this area. It is noteworthy that in all our eight cases of hyperextension of the cervical spine, the lower cervical cord segments were indeed involved.

Flexion Injuries

In 1907, Lloyd 54 was the first to suggest flexion instead of extension as the principal mechanism involved in cord compression without vertebral fracture or deformity. In his "flexion-recoil" theory, he rea- soned that after forward displacement, producing the neural injury, the upper cervical segment sprang back as a result of muscle action to give a normal appearance on x-ray films. Barnes in 194811 categorically refuted Lloyd's flexion-recoil theory: his studies on adult cadavers showed that for flexion dislocation to damage the cord, there would have to be either fracture of the facet joints or unilateral locking of at least one facet, which would make the forward displacement irreducible spontaneously. In either case, the x-ray film would be abnormal.

Since Barnes' report, a number of researchers have noted flexion as the culpable mechanism in cases of pediatric spinal cord injury without vertebral damage. The four cases of Burke, z4 two cases of Glasauer and


Cares, 41,4z and one case of Cheshire 31 all had complete cord transection following a flexion injury. Moreover, Teng and Papatheodorou, 73 Dunlap, et aL, 34 and Pa- pavasiliou 6~ have all described the syndrome of traumatic flexion subluxation of the cervical spine during childhood with normal radiography. Although Barnes' refutation of Lloyd's "flexion-recoil" theory might be entirely justified in the adult, it is conceivable that flexion dislocation could spontaneously reduce itself without facet fracture or locking in the highly supple and mobile spine of a child.

The following anatomical features in the pediatric cervical spine account for its increased physiologic mobility as well as its susceptibility to flexion injury: 1) The interspinous ligaments, posterior joint capsule, and cartilaginous end plates are elastic and often redundant. 9,69,76 2) The articulating surfaces of the facet joints are more horizontally oriented. 9,3~ 3) The anterior portion of the vertebral bodies is wedged forward so that anterior slipping between adjacent bodies is facilitated. 9,69 4) In the mature spine, the two uncinate processes project upward and outward to articulate with the corresponding lower borders of the body above at the uncovertebraljoints. The characteristic orientation of the fully developed uncinate processes normally limits lateral and rota- tional movements between adjacent bodies. In infants and children under 10 years, the uncinate processes are flat and therefore ineffective in withstanding flexion-rotation forces. 19,74 5) The proportionately heavy head and relatively underdeveloped musculature of the infant neck constitute an unusual susceptibility for flexion-extension injuries, z7'28'68,75

Large-scale anatomical studies of normal children reveal that the horizontal orientation of the facet joints as well as the anterior wedging of the vertebral bodies is much more prominent in the upper three to four segments of the cervical spine2 ,69 Also, according to Townsend and Rowe, TM Baker and Berdon, 1~ and Braakman and Penning, TM the fulcrum for maximum flexion in young children is at C2-3 and C3-4, whereas the fulcrum for maximum flexion in the adult is at C5-6. It is therefore not surprising that the upper cervical segments in children display the greatest physiological mobility, 3~ and that flexion subluxation in infants and younger children most often involves the upper segments of the cervical cord. 47,60,73,74 In the present series, eight children had flexion compression of the cord and in five the cord segments injured were between C-I and C-4. In only three children was the lower cervical cord involved.

According to Bailey 9 and von Torklus and Gehle, TM

the characteristic anatomical features of the young spine gradually approach adult status by 8 years of age. Thus, the anterior wedging of the bodies disap- pears, the articulating planes of the facets become more vertical, the uncinate processes gain height, and the ligaments and capsules increase in tensile strength. It follows that the cervical spine, especially its upper


11


portion, is at maximum risk for forward slipping against flexion forces during the first few years of life. As the child develops beyond 8 years, the cervical spine gains resistance against flexion insults, and by 16 to 18 years, when the adult status is reached, flexion forces will more likely produce the well known syndrome of fracture or fracture-subluxation. This is, indeed, borne out by our data: seven out of eight cases of flexion injuries involved children under 8 years (Table 2), with five involving the C1-4 segments (Table 6). Cord injuries due to flexion are also more serious in the younger children, again reflecting the inherent instability of the spine at this age: the six children with severe or complete cord syndromes were all under 8 years of age (Table 5 and Fig. 1).

The case of the abused child with severe central cord syndrome represents a special instance of repetitive flexion and hyperextension injury. 27,28,7~ During the paroxyms of shaking in the anteroposterior plane, the heavy infantile head supported poorly by weak neck muscles courses through a two-phase cycle of rapid, repetitive flexion of the head until the chin strikes the anterior chest, alternating with extensions of the head until the occiput strikes the back. Al- though the elastic spinal column escapes fracture- dislocation, the underlying cord necessarily suffers multiple battering.

Longitudinal Distraction Injuries

In 1974, Burke 24 described autopsy findings of a constricted segment of the cervical cord 4 cm in length in an 11-month-old infant. In another infant with a T-4 paraplegia, he found during laminectomy a similarly narrowed and elongated segment of traumatized cord unlike the usual discrete, short segment seen after compression injury. Glasauer and Cares 41,42 reported similar findings in two other infants, one at autopsy and the other at surgery. They postulated that longitudinal distraction rather than flexion or extension-compression was the mechanism involved in these infants.

In this context, Leventhal's ~3 cadaver study of the longitudinal compliance of the neonatal spine is most revealing. He found that the elastic spinal column of the neonate can be stretched 2 in. without signs of structural disruption, but the spinal cord, devoid of elastic elements, can only stretch �88 in. before ruptur- ing. During a forceful breech extraction, the spinal cord can be ruptured by longitudinal distraction, along with its investing dura and leptomeninges, whereas the vertebral column will remain completely intact. 1,7~ None of Leventhal's six cases had any evidence of radiographic abnormalities.

It is reasonable to assume that the younger the victim, and therefore, the more elastic the spine, the more serious the myelopathy will be with longitudinal distraction. In our series, there were two cases of longitudinal distraction, both involving very young

children. Both received serious damage to the cord caused by extremely violent forces; one had a lower cervical cord injury and the other a complete lesion at T-3.

Ischemic Injuries

At least two cases of spinal cord infarction following minor trauma in children can be found in the literature. Ahmann, et al.,Z described the autopsy findings in a 4-year-old child who sustained a relatively trivial hyperextension injury to the neck and who presented with a high cervical cord lesion of delayed onset. Bilateral gray matter infarction was found from C-3 to T-2. A 22-month-old child, also with a delayed cervical cord lesion following minor hyperextension, was also found to have infarction of the dorsolateral columns from the cervicomedullary junction to T-5. Both children had normal plain films and myelograms. Ahmann, et aL, 2 postulated that the vertebral arteries could have been temporarily occluded or thrown into spasm at the time of hyperextension, and longitudinal watershed infarction would result if col- lateral flow from thoracic and lower cervical medullary arteries was insufficient. He added that if surface vessels in the dorsal coronary plexus were compressed at multiple levels during hyperextension, infarction could occur within the ventrodorsal watershed area between terminal supplies from the dorsal plexus and the anterior sulcal arteries.

Gilles, et al.,4~ made a detailed anatomical study of the infantile atlanto-occipital junction that has shed much light on the subject of the vulnerability of the infantile vertebral arteries during trauma. They found the infantile atlanto-occipital joint to be inherently unstable. The small arch of C- 1 resting against a large foramen magnum, the weak and redundant alar, apical, and atlanto-occipital ligaments, the elastic and lax joint capsules investing the occipital condyles, and the condyles' own flattened surface, all contribute to an inherent instability in this region where flexion and extension customarily induce horizontal sliding movements. Moreover, the lateral mass of C-1 and the posterior portion of the occipital condyle, which normally form the protective groove for the vertebral artery as it curls behind the lateral mass to enter the skull, are both stunted in height. This exposes the artery to ready compression by the to-and-fro movements between the condyle and the C-1 arch during hyperextension. Gilles, et al.,4~ actually demonstrated bilateral occlusion of the vertebral arteries with the neck in extension at postmortem angiography. If this were an antemortem event, the upper cervical cord could be rendered ischemic.

A final important point concerning ischemic injury is that six patients in our series had severe hypotension on admission. A suboptimal perfusion pressure to the traumatized cord with impaired autoregulation of blood flow could have contributed to the final insult.


12

Other Hypotheses

In 1915, Holmes 47 postulated that a direct blow to the vertebral column insufficient to cause skeletal damage or deformity could set up shock-wave oscil- lations in the subjacent cord at a frequency different from that in the column of bone, causing "slapping damage" to the cord against the bone wall of the central canal. This has never been substantiated.

Cramer and McGowan in 194432 suggested that during flexion, the intervertebral disc protrudes backward by means of a hydraulic piston-like action of the nucleus pulposus, which then spontaneously retracts back inside the anulus. We now know that traumatic disc extrusion with rupture of the anulus does occur rarely in adults but usually is associated with fracture of the adjacent vertebral bodies. The relevance of this to the pediatric syndrome of SCIWORA is uncertain, but myelography may be helpful in its exclusion in doubtful situations.

The Missed Occult Fracture

On rare occasions, the physician may be misled into making the diagnosis of SCIWORA if an occult fracture is missed on conventional radiography. This can be hazardous, for, as the physician is busy searching for esoteric mechanisms to explain the spinal cord injury in the absence of fracture-dislocation, the patient's spine remains dangerously unstable. Marar ~6 demonstrated in 12 autopsy cases that a horizontal fracture of the vertebral body below the pedicles can occur in the cervical spine with hyperextension. The anterior longitudinal ligament is always torn, but reflex muscle action may temporarily reduce this fracture so perfectly that visualization with conventional plain films may be impossible. None of these 12 patients had abnormal antemortem x-ray films. Re- duction in such cases is temporary; instability will become obvious when reflex muscle spasm subsides. Vines 77 also encountered lateral mass fractures not extending to the facet surface missed by routine radiographic views. Thus, multiplane tomographic studies or CT should be employed to rule out these occult fractures.

The Neurological Lesions

It is obvious from our figures that the neurological lesion in the pediatric SCIWORA syndrome is more serious than the type of neurological lesions found in adult patients with spinal cord injury and normal x-ray films. Most authors agree that the mechanism for the adult patients is one of hyperextension super- imposed on preexisting cervical spondylosis, and the neurological picture is predominantly that of acute central cord syndrome, a,8,'2'21'38,4~'55'5~'66'67 Bedbrook 13'14 and Schneider, et al., 66 have furnished pathological confirmation of the prevalence of central cord necrosis in adult hyperextension injury. Neurological lesions other than central cord syndrome are uncommon.


Bedbrook mentioned only two cases of anterior cord syndrome, two cases of Brown-Srquard syndrome, and four cases of complete cord lesions in a group of 63 adults with hyperextension injury and normal radiography. 13 This probably explains the generally fa- vorable long-term prognosis for these patients, for the central cord syndrome is usually associated with good recovery, whereas complete cord transection and anterior cord syndromes have a more sinister prognosis. 15,25,45 In our series of pediatric SCIWORA, seven children had complete transections and seven others had severe incomplete cord syndromes, representing an incidence of 58.3% with serious spinal cord damage. The data from others, although scanty, reflect the same distressing outlook for children with SCIWORA.24,31,44,65

Although children with SCIWORA have worse prognoses than adult patients with hyperextension injury and normal radiography, they fare consider- ably better than those children whose spinal cord injuries are associated with fractures or dislocations. Twelve such children were treated at our institution from 1960 to 1980; two died, eight had complete transections, one had a severe incomplete lesion, and only one had mild neurological injury. None of the children with initially complete transections made any significant neurological recovery. These figures con- cur with other series on pediatric spinal cord trauma with bone injuries: Burke 24 reported 86% of cases with complete cord lesions; Kewalramani, et aL, ~~ 60%; Hubbard 48 57.1%; Hachen 44 88.9%; and Scher 65 85.7%.

Our data also indicate several important differences in the neurological status between children younger than 8 years and those over 8 years. The neurological injuries encountered in children younger than 8 years are much more serious than those in the older children. All the complete transections and all but one of the severe incomplete lesions were found in children 6 months to 8 years, whereas all but one older child had mild lesions (Fig. 1). The same age influence is also apparent within the subgroup with central cord syndrome: the children with more severe lesions are clearly younger than those with mild lesions (Fig. 1). This supports the observation that the inherent instability of the pediatric spine is maximum in infancy and decreases as the child reaches the second decade of life. The pattern of neurological lesions in children from 8 to 16 years is not unlike that seen in the adult counterpart of SCIWORA.

Upper cervical cord injuries are also more common in the younger age group. In contrast, lower cervical cord injuries occur in equal frequency throughout the age range of the series (Fig. 2). This is due to the fact that the physiological hypermobility peculiar to infants and young children predominantly involves the upper two to three cervical segments and not the cervicothoracicjunction as much as in adults. Most of the age-related changes in anatomy and biomechanics occur in the upper segments, while the lower segments


13


go through a much more subtle transition between the childhood and adult status.

Marar 57 concluded from his study of 126 patients with cervical spine injuries and cord damage that the mechanism of the injury can be predicted by the neurological lesions. The suggested mechanisms are: 1) flexion dislocation for complete cord transection; 2) hyperextension for central cord syndrome; 3) burst fracture or disc retropulsion for anterior cord syndrome; and 4) unilateral facet subluxation for Brown- Srquard syndrome. Unlike Marar, we found no correlation between the mechanism of injury and the type or completeness of the neurological lesions in children with SCIWORA. Certainly, the central cord syndrome characteristically associated with hyperextension in the adult is as often caused by flexion as by hyperextension in these children. Also, complete cord transection can be a result of hyperextension, flexion, longitudinal distraction, or direct crush injury. Our one case of Brown-Srquard syndrome was caused by hyperextension. Similarly, the unclassifiable partial syndrome can be caused by either flexion or hyperextension. Although it appears from Table 5 that flexion more often causes severe or complete lesions and that hyperextension more often causes mild lesions, the distinction is spurious, for the more serious neurological injuries all involved infants or young children. The younger the child, the more deformable the spine and the worse the neural insult, regardless of the mode or the direction of deformation.

Delayed Onset of Neurological Manifestation

The phenomenon of delayed onset of neurological manifestation in children with SCIWORA has been described previously. Cheshire 31 reported three patients with this delay, Burke 23 reported one, and Ah- mann, et al., 2 two infants with similar delay and slow evolution of neurological signs. The paralysis developed rapidly once it began in all six children and culminated in complete cord lesions. In our series, delayed onset seemed to be the predominant mode of clinical presentation.

We have no ready explanation for this delay, but several points are worthy of note. Seven patients vividly recalled transient neurological symptoms at the time of injury. These include paresthesia, numbness, lightning sensation, and a subjective feeling of "total body paralysis." If the spinal cord injury in SCIWORA is caused by some form of bone compression secondary to self-reducing subluxation, these symptoms must arise at the exact time of contact of bone with neural tissues. Because of rapid spontaneous reduction, complete neural destruction does not occur immediately. The severe neurological damage following the "latent period" must be a result of one of two occurrences: 1) incipient instability developed at the time of the original subluxation, which was then reactivated repeatedly by continued movements of the spine, thereby causing multiple repetitive insults to

the cord; or 2) the original injury to the cord set off slow but progressive destruction of neural tissues.

If the first tenet is correct, the original subluxation must have resulted in partial tearing of crucial ligaments normally responsible for stability. Although immediate muscle action resulted in spontaneous reduction, such ligamentous injury will permit significant but not easily demonstrable displacements with each normal cycle of flexion-extension, causing a form of "punch drunk" trauma to the cord. Scher, 62 Evans, 35 and Webb, et aL, 79 all alluded to this type of damage to the posterior ligaments in their so-called "hidden flexion injury of the cervical spine." This mechanism is strongly suggested by the case of the child (Case 3 in Table 7) who experienced the initial "feeling of paralysis" but remained well until he played basketball 2 days following his neck injury. It may also be involved in the child (Case 4 in Table 7) who began recovering from his central cord syndrome while wearing a cervical collar but then experienced recurrence of paralysis after he removed his collar during play. These two were the only children with reported incidents involving strenuous neck movements during the latent period. Radiographic confirmation of this incipient instability is lacking since none of these 12 children had abnormal dynamic films. However, these films were made after, and not before, the onset of neurological signs, and the attend- ant spasm could have masked whatever incipient instability that was present before the cord was per- manently injured. All of these patients had late (1 to 2 years) x-ray films to rule out delayed spinal deformities, but none was ever found.

The second tenet implies a slow but relentlessly progressive insult to the cord following a single contact or stretch injury. This insult may be ischemia. Ahmann, et al., 2 reported two infants with delayed onset; both had extensive watershed infarctions of the cord at autopsy, and the authors linked this lesion to the syndrome of progressive or evolving cerebral stroke. We know of no other case of proven cord infarction secondary to S C I W O R A . A slowly en- larging epidural hematoma may cause gradual compression, but Bedbrook a3 categorically denied the existence of such an entity following spinal injury. Certainly, none of the myelograms performed on our patients, including six cases with delayed evolution of signs (Table 7), showed a subarachnoid block. Other possibilities are delayed traumatic hematomyelia, progressive edema, or the central hemorrhagic necrosis claimed by Osterholm ~9 to be caused by accumulation of putative amines. None of these hypotheses have been proven in our patients.

We were unable to detect any unique clues in the age distribution, mechanisms of injury, or radiographic features of this subgroup to distinguish it from those patients with no delay in onset. The pattern of neurological outcome in this subgroup is also comparable to that of the main series: the complete lesions

J. Neurosurg. / Volume 57 / July, 1982 125

14

remained complete; those with initially severe lesions remained severely disabled; and those with initially mild lesions made generally satisfactory recovery. It is noteworthy that once sensorimotor paralysis began at the end of the latent period, it progressed inexorably into its established state within hours. The importance of this subgroup, therefore, rests on the hope that if the early warning signs of transient symptoms could be recognized and promptly acted upon before the onset of neurological signs, the tragic fate of some of these children might be duly averted.

Complications

The early and late complications of the pediatric SCIWORA syndrome are the same as those found in cord-injured children with fracture-dislocation, with one exception: delayed and progressive spinal deformities such as scoliosis, kyphosis, and lordosis were unknown in our series through a follow-up period of 18 years. Such deformities are common among children with skeletal injuries of the spine. Campbell and Bonnett 29 reported an incidence of 91% in their children and Burke 24 reported a 55.2% incidence. Half of Hubbard's series 48 had some degree of spinal deformity, especially those with initially unstable spines. Babcock s suggested that the late deformities are a result of a combination of factors, including destruction of the growth centers in the centrum, ischemic necrosis of epiphyseal growth plates, unilateral bone loss or wedging, unilateral fusion between fractured vertebral segments, and imbalance of the paraspinous muscle. It may be assumed that, since the first four factors all involve bone injuries in the premature spine and are therefore not relevant to the SCIWORA syndrome, paraspinous muscle imbalance alone is not sufficient to cause delayed deformities.

Outcome

For children with SCIWORA the long-term prognosis is poor. Neither age nor treatment affect the final outcome. The only prognosticating factor is the initial neurological status. The trend of neurological recovery follows a grim but consistent pattern: useful recovery does not occur in those with initially complete lesions and seldom occurs in those with severe incomplete lesions. The best outcome is in those patients with initially mild deficits; most will make a complete recovery, and the rest will be only minimally disabled. The level of the neurological lesion, however, does influence the rehabilitation potential of the victim. Those with higher lesions generally do poorly compared to those with thoracic lesions. All of the high cervical lesions in this series occurred in infants or young children. Since quadriplegic children who are also ventilator-dependent have very limited rehabilitative potentials, these young children with SCIWORA represent a group with a very low rate of re-entry to the community.

D. P a n g a n d J. E. Wi lbe rge r , Jr.

Management

The diagnosis of SCIWORA should only be made after occult fractures have been excluded by polyto- mography or CT scanning. Occasionally, myelography may be necessary to rule out traumatic disc extrusion or extradural hematoma, but none of the 12 myelograms in our series showed a subarachnoid block.

Furthermore, gross instability must be ruled out by flexion-extension films. Although the initial subluxation had been reduced by spontaneous muscle action, there could have been sufficient stretch injury to the ligaments to render the spine unstable. Cheshire 31 cited one case of instability missed by plain films and tomography but discovered by dynamic radiographic studies. Such studies must be executed under careful control. Ideally, the patient should be awake and cooperative so that continued neurological examination can be done to monitor the status of the cord. Fluoroscopy should be used, and troublesome paraspinal muscle spasm may be partly neutralized by muscle relaxant. The spasm, although serving an important protective function, interferes with obtaining good flexion-extension films. In some of the cases of Webb, et al., 79 the spasm virtually precluded any active flexion, and subtle instability could not be ruled out until several days later. An inadequate dynamic study should never be accepted as proof of stability.

If immediate instability is revealed by the flexion- extension study, either surgical fusion or halo fixation is recommended. There are often objections to using only external immobilization for pure ligamentous injuries for fear of nonhealing, but since in the pediatric spine the ligaments are probably not ruptured but stretched, we think that halo fixation is a viable option.

If no immediate instability is demonstrated, or if spasm prevents adequate flexion, the cervical spine should be immobilized by a well fitting stiff collar. We find this adequate immobilization for this group of children with ligamentous sprain. In 5 to 7 days, or when the initial spasm has subsided, delayed dynamic films should be obtained. Both Webb, et aL, TM and Scher 62 have mentioned patients whose instability was discovered only on delayed dynamic films. One child in our series was discovered to have delayed anterior subluxation on the late dynamic films.

We urge that children who present with seemingly trivial head and neck injuries be questioned specifi- cally for transient neurological symptoms. If these are present, tomography and dynamic films should be obtained, followed by admission for observation.

Summary and Conclusions

1. The pediatric syndrome of spinal cord injury without radiographic abnormality (SCIWORA) includes children with traumatic myelopathy who have


15


no radiographic evidence of fracture or dislocation on initial examination.

2. The mechanisms include hyperextension, flexion, repetitive flexion-extension, longitudinal distraction, and direct crush injury. Anatomical features peculiar to children permit transient subluxation without bone injury. Reflex muscle action causes spontaneous reduction which gives the normal radiographic appearance.

3. The neurological lesions encountered included complete cord syndrome, Brown-S&tuard syndrome, central cord syndrome, and partial cord syndrome that did not fit the central or anterior cord patterns. Fifty-eight percent of these lesions were severe or complete,

4. There were differences in the neurological presentations of children younger than 8 years compared to children over 8 years. The younger children had more serious neural injuries and a higher incidence of upper cervical cord and thoracic cord injuries. Within the subgroup of central cord syndrome, the younger children also suffered more severe injuries.

5. Fifty-two percent of children had delayed onset of paralysis, and among these many recalled transient symptoms including paresthesia, numbness, and a subjective feeling of paralysis. All children with head and neck injuries complaining of symptoms must be fully investigated.

6. Initial radiographic examination should include tomography to rule out occult fractures and occasionally myelography to rule out traumatic disc extrusion. Immediate dynamic films should be obtained to select those children with incipient instability, who may then require surgical fusion or immobilization with halo fixation. Delayed dynamic films should be obtained to rule out late instability.

7. Long-term prognosis of SCIWORA is poor. Children with complete lesions and most of those with severe lesions do not recover. Only those with initially mild lesions have hope for satisfactory recovery. The initial neurological status is the major factor that determines the extent of long-term recovery.

References

1. Abroms IF, Bresnan MJ, Zuckerman JE, et al: Cervical cord injuries secondary to hyperextension of the head in breech presentations. Obstet Gynecoi 41:369-378, 1973

2. Ahmann PA, Smith SA, Schwartz JF, et al: Spinal cord infarction due to minor trauma in children. Neurology 25:301-307, 1975

3. Alexander E Jr, Davis CH Jr, Field CH: Hyperexten- sion injuries of the cervical spine. Arch Neuroi Psychia- try 79:146-150, 1958

4. Anderson MJ, Schutt AH: Spinal injury in children. A review of 156 cases seen from 1950 through 1978. Mayo Clin Proc 55:499-504, 1980

5. Andrews LG, Jung SK: Spinal cord injuries in children

in British Columbia. Paraplegia 17:442--451, 1979 6. Audic B, Maury M: Secondary vertebral deformities in

childhood and adolescence. Paraplegia 7:11-16, 1969 7. Aufdermaur M: Spinal injuries in juveniles. Necropsy

findings in twelve cases. J Bone Joint Sorg (Br) 56: 513-519, 1974

8. Babcock JL: Spinal injuries in children. Pediatr Clin North Am 22(2):487-500, 1975

9. Bailey DK: The normal cervical spine in infants and children. Radiology 59:712-719, 1952

10. Baker DH, Berdon WE: Special trauma problems in children. Radiol Clin North Am 4:289-305, 1966

11. Barnes R: Paraplegia in cervical spine injuries. J Bone Joint Surg (Br) 30:234-244, 1948

12. Bedbrook GM: Compression, flexion and extension injuries of the cervical spine with tetraplegia. Proc Veterans Adm Spinal Cord lnj Conf 19:6-28, 1973

13. BedbrookGM: Closed injuries ofthe cervicalspine and spinal cord extension-rotation injuries. Proc Veterans Adm Spinal Cord Inj Conf 19:.58-59, 1973

14. Bedbrook GM: Pathological principles in the management of spinal cord trauma. Paraplegia 4:43-56, 1966

15. Bedbrook GM: Spinal injuries with tetraplegia and paraplegia. J Bone Joint Snrg (Br) 61:267-284, 1979

16. Berkin CR, Hirson C: Hyperextension injury of the neck with paraplegia. J Bone Joint Surg (Br) 36:57-6 l, 1954

17. Bourmer HR: Zur Frage der Halsmarksch~idigung bei Hyperextensionsverletzungen der Wirbels~iule. Arch Kiln Chir 268:409--416, 1951

18. Braakman R, Penning L: The hyperflexion sprain of the cervical spine. Radiol Clin 37:309-320, 1968

19. Braakman R, Penning L: Injuries of the Cervical Spine. Amsterdam: Excerpta Medica, 1971, p 20

20. Braakman R, Penning L: Injuries of the Cervical Spine. Amsterdam: Excerpta Medica, 1971, p 75

21. Braakman R, Penning L: Mechanisms of injury to the cervical cord. Paraplegia 10:314-320, 1973

22. Breig A, E1-Nadi AF: Biomechanics of the cervical spinal cord. Relief of contact pressure on and over- stretching of the spinal cord. Acta Radioi (Diag) 4: 602-624, 1966

23. Burke DC: Spinal cord trauma in children. Paraplegia 9:1-14, 1971

24. Burke DC: Traumatic spinal paralysis in children. Paraplegia 11:268-276, 1974

25. Burke DC, Murray DD: The management of thoracic and thoraco-lumbar injuries of the spine with neurological involvement. J Bone Joint Surg (Br) 58:72-78, 1976

26. Byers RK: Transection of the spinal cord in the new- born. A case with autopsy and comparison with a normal cord at the same age. Arch Nenrol Psychiatry 27:585-592, 1932

27. Caffey J: On the theory and practice of shaking infants. Its potential residual effects of permanent brain damage and mental retardation. Am J Dis Child 124:161-169, 1972

28. Caffey J: The whiplash shaken infant syndrome: man- ual shaking by the extremities with whiplash-induced intracranial and intraocular bleedings, linked with residual permanent brain damage and mental retardation. Pediatrics 54:396-403, 1974

29. Campbell J, Bonnett C: Spinal cord injury in children. Clin Orthop 112:114-123, 1975


16

30. Cattell HS, Filtzer DL: Pseudosubluxation and other normal variations in the cervical spine in children. A study of one hundred and sixty children. J Bone Joint Surg (Am) 47:1295-1309, 1965

31. Cheshire DJE: The pediatric syndrome of traumatic myelopathy without demonstrable vertebral injury. Paraplegia 15:74-85, 1977

32. Cramer F, McGowan F J: The rrle of the nucleus pulposus in the pathogenesis of so called "recoil" injuries of the spinal cord. Surg Gynecoi Obstet 79: 516-521, 1944

33. Crooks F, Birkett AN: Fractures and dislocations of the cervical spine. Br J Surg 31:252-265, 1944

34. Dunlap JP, Morris M, Thompson RG: Cervical-spine injuries in children. J Bone Joint Surg (Am) 40: 681-686, 1958

35. Evans DK: Anterior cervical subluxation. J Bone Joint Surg (Br) 58:318-321, 1976

36. Forni I: Le fratture del rachide nel bambino. Chir Org Movimento 31:347-361, 1947

37. Forsyth HF: Extension injuries of the cervical spine. J Bone Joint Surg (Am) 46:1792-1797, 1964

38. Frankel HL, Hancock DO, Hyslop G, et al: The value of postural reduction in the initial management of closed injuries of the spine with paraplegia and tetraplegia. I. Paraplegia 7:179-192, 1969

39. Gelehrter G: [Fracture of the vertebrae in children and adolescents.] Arch Orthop Unfallchir 49:253-263, 1957 (Get)

40. Gilles FH, Bina M, Sotrel A: Infantile atlantooccipital instability. The potential danger of extreme extension. Am J Dis Child 133:30-37, 1979

41. Glasauer FE, Cares HL: Biomechanical features of traumatic paraplegia in infancy. J Trauma 13:166-170, 1973

42. Glasauer FE, Cares HL: Traumatic paraplegia in infancy. JAMA 219:38-41, 1972

43. Gordon N, Marsden B: Spinal cord injury at birth. Neuropaediatrie 2:112-118, 1970

44. Hachen HJ: Spinal cord injury in children and adolescents: diagnostic pitfalls and therapeutic considerations in the acute stage. Paraplegia 15:55-64, 1977

45. Hardy AG: The management of paraplegia due to fracture dislocations of the dorsolumbarjunction. Proc Veterans Adm Spinal Cord Inj Conf 19:145-150, 1973

46. Hasue M, Hoshino R, Omata S, et al: Cervical spine injuries in children. Fukushima J Med Sci 20:115-123, 1974

47. Holmes G: Spinal injuries of warfare. Br Med J 2: 769-774, 1915

48. Hubbard DD: Injuries of the spine in children and adolescents. Clin Orthop 100:56-65, 1974

49. Hughes JT, Brownell B: Spinal-cord damage from hyperextension injury in cervical spondylosis. Lancet 1:687-690, 1963

50. Kewalramani LS, Kraus JF, Sterling HM: Acute spinalcord lesions in a pediatric population: epidemiological and clinical features. Paraplegia 18:206-219, 1980

5 I. Koch BM, Eng GM: Neonatal spinal cord injury. Arch Phys Med Rehabil 60:378-381, 1979

52. Kraus JF: Epidemiological features of head and spinal cord injury, in Schoenberg S (ed): Neurological Epide- miology: Principles and Clinical Applications. Advances


in Neurology, Vol 19. New York: Raven Press, 1978, pp 261-279

53. Leventhal HR: Birth injuries of the spinal cord. J Pediatr 56:447-453, 1960

54. Lloyd S: Fracture dislocation of the spine. Med Rec 71:465-470, 1907

55. Logue V: Cervical spondylosis, in Williams D (ed): Modern Trends in Neurology, 2nd Series. London: Butterworths, 1957, pp 259-273

56. MararBC: Hyperextensioninjuriesofthecervicalspine. The pathogenesis of damage to the spinal cord. J Bone Joint Surg (Am) 56:1655-1662, 1974

57. Murat BC: The pattern of neurological damage as an aid to the diagnosis of the mechanism in cervical-spine injuries. J Bone Joint Surg (Am) 56:1648-1654, 1974

58. Melzak J: Paraplegia among children. Lancet 2:45-48, 1969

59. Osterholm JL: The vascular and cellular basis for spinal cord hemorrhagic necrosis, in Morley TP (ed): Current Controversies in Neurosurgery. Philadelphia: WB Saunders, 1976, pp 100-109

60. Papavasiliou V: Traumatic subluxation of the cervical spine during childhood. Orthop Clin North Am 9(4): 945-954, 1978

61. Robson PN: Hyperextension and haematomyelia. Br Med J 2:848-852, 1956

62. Scher AT: Anterior cervical subluxation: an unstable position. A JR 133:275-280, 1979

63. Scher AT: Cervical spinal cord injury without evidence of fracture or dislocation. An assessment of the radiological features. S Afr Med J 50:962-965, 1976

64. Scher AT: Diversity of radiological features in hyperextension injury of the cervical spine. S Afr Med J 58:27-30, 1980

65. Scher AT: Trauma of the spinal cord in children. S Afr Med J 50:2023-2025, 1976

66. Schneider RC, Cherry G, Patek H: The syndrome of acute central cervical spinal cord injury. With special reference to the mechanisms involved in hyperextension injuries of cervical spine. J Neurosurg 11:546-577, 1954

67. Schneider RC, Thompson JM, Bebin J: The syndrome of acute central cervical spinal cord injury. J Neurol Neurosurg Psychiatry 21:216-227, 1958

68. Sherk HH, Schut L, Lane JM: Fractures and dislocations of the cervical spine in children. Orthop Clin North Am 7(3):593-604, 1976

69. Sullivan CR, Bruwer AJ, Harris LE: Hypermobility of the cervical spine in children: a pitfall in the diagnosis of cervical dislocation. Am J Surg 95:636-640, 1958

70. Swischuk LE: Spine and spinal cord trauma in the battered child syndrome. Radiology 92:733-738, 1969

71. Taylor AR: The mechanism of injury to the spinal cord in the neck without damage to the vertebral column. J Bone Joint Surg (Br) 33:543-547, 1951

72. Taylor AR, Blackwood W: Paraplegia in hyperextension cervical injuries with normal radiographic appear- ances. J Bone Joint Surg (Br) 30:245-248, 1948

73. Teng P, Papatheodorou C: Traumatic subluxation of C2 in young children. Bull Los Angeles Neurol Soc 32: 197-202, 1967

74. Tondury G: The cervical spine: its development and changes during life. Acta Orthop Belg 25:602-607, 1959

75. Towbin A: Spinal injury related to the syndrome of

128 J. Neurosurg. / Volume 57/July, 1982

17


sudden death ("crib-death") in infants. Am J Clin Pa- thol 49:562-567, 1968

76. Townsend EH Jr, Rowe ML: Mobility of the upper cervical spine in health and disease. Pediatrics 10: 567-573, 1952

77. Vines FS: The significance of "occult" fractures of the cervical spine. A JR 107:493-504, 1969

78. Von Torklus D, Gehle W: The Upper Cervical Spine. New York: Grune and Stratton, 1972, pp 10-94

79. Webb JK, Broughton RBK, McSweeney T, et al: Hid-

den flexion injury of the cervical spine. J Bone Joint Surg (Br) 58:322-327, 1976

Manuscript received November 19, 1981. Accepted in final form February 22, 1982. Address reprint requests to: Dachling Pang, M.D., Chil-

dren's Hospital of Pittsburgh, 125 DeSoto Street, Pittsburgh, Pennsylvania 15213.

J. Neurosurg. / Volume 57 / July, 1982 129

18

Pediatric Traumatic Atlanto-Occipital Dislocation: Five Casesand a Review

*Keith Kenter, M.D., ‡Gordon Worley, M.D., †Tish Griffin, R.N., and †‡Robert D. Fitch, M.D.

Study conducted at Duke University Medical Center, Durham, North Carolina, U.S.A.

Summary: Traumatic atlanto-occipital dislocation (AOD) hasbeen thought to be a rare and fatal injury. Recently, moresurvivors, especially children, have been reported. During a10-year period, the authors have encountered five children withtraumatic AOD. A retrospective review of traumatic AOD inchildren from 1985 to 1995 was performed. Clinical presenta-tion, initial radiologic findings, and final outcome were empha-sized. Distance from the dens to the basion and the ratio ofPowers were measured from initial lateral cervical spine radio-graphs. The average distance from the dens to the basion was

9.8 mm. The average ratio of Powers was 1.38. There werethree survivors, two having a concomitant spinal cord injury.All survivors underwent a posterior occipitovertebral fusion.Three cases initially went undiagnosed. The diagnosis of AODby lateral cervical spine radiographs can be difficult. The au-thors recommend detailed measurements of the initial cervicalspine radiographs in pediatric patients at risk for traumaticAOD. Key Words: Atlanto-occipital dislocation—Spinal cordinjury—Pediatric—Cervical spine measurements.

In the past, traumatic atlanto-occipital dislocation(AOD) was considered a rare and usually fatal injury(6,17). Recently, more survivors, particularly children,have been reported, perhaps as a result of quicker re-sponse time to the scene of accidents by emergencymedical teams, better initial cervical spine immobiliza-tion, and earlier diagnosis of AOD in the emergencydepartment (10,16,17,19,21,27).

Blackwood (2) reported in 1908 the first case of trau-matic AOD. Postmortem examinations of patients whodied of “multiple trauma” documented an overall preva-lence of AOD of 8%; however, the prevalence of patientswho died of brain and spinal cord injuries was greater(20%–25%) (1,3,6). Bulas et al. (4) found an 0.7% inci-dence of pediatric traumatic AOD in 1,600 children dur-ing a 5-year period.

The diagnosis of traumatic AOD is dependent on bothrecognizing the physical findings of cervical spine in-jury, including cervical spinal cord injury, in the pres-ence of head trauma and on properly interpreting imag-ing studies. Because prompt diagnosis may improve out-come, we reviewed our cases of traumatic AOD over a10-year period. The purpose of this study was to report

our experience from a major level 1 trauma institution ontraumatic AOD in the pediatric population.

MATERIALS AND METHODS

A retrospective review of children with traumaticAOD from 1985 to 1995 was performed. Pediatric pa-tients diagnosed at discharge with traumatic AOD wereidentified by performing a computer search of the DukeUniversity Medical Center (DUMC) Medical RecordsDatabase. During this 10-year period we identified fivechildren with traumatic AOD. Charts were then reviewedand pertinent clinical information was recorded. Thecross-table lateral cervical spine radiographs were ob-tained from the Department of Radiology at DUMC orfrom the outside hospital where the patient was origi-nally seen and transferred to DUMC. Measurements de-scribed by Wholey et al. (26) and Powers et al. (20) weremade from these initial radiographs. Wholey et al. mea-sured the distance between the tip of the dens and theoverlying basion (DB distance) to determine whethertraumatic AOD was present (Fig. 1). A DB distance of!5 mm is considered normal in adults, and !10 mm isnormal in children. Powers et al. used the ratio of thedistance from the basion to the posterior arch of the atlasdivided by the distance from the opisthion to the anteriorarch of the atlas to determine whether traumatic AODwas present (Fig. 2). A ratio of >1.0 is considered diag-nostic of traumatic AOD. Figure 3 shows a typical lateralcervical spine radiograph with our measurements fromone of our described patients.

Address correspondence and reprint requests to Keith Kenter, M.D.,Assistant Professor Department of Orthopaedic Surgery, University ofMissouri, One Hospital Drive, Columbia, MO 65212, U.S.A. E-mail:[email protected]

From the *Department of Orthopaedic Surgery, University of Mis-souri, Columbia, Missouri, and the ‡Department of Pediatrics and †Di-vision of Orthopaedic Surgery, Duke University Medical Center,Durham, North Carolina, U.S.A.

Journal of Pediatric Orthopaedics21:585–589 © 2001 Lippincott Williams & Wilkins, Inc., Philadelphia

585

19

RESULTS

Table 1 summarizes the clinical features of our fivepediatric patients with traumatic AOD diagnosed duringa 10-year period at DUMC. There were three girls andtwo boys with an average age of 8 years (range 7–12).Three children were unrestrained passengers in a motorvehicle crash and two children were pedestrians struckby an automobile. An associated closed head injury waspresent in four of the five children. These four patientsall had loss of consciousness and were intubated andmechanically ventilated. Facial injuries were present infour of the five children and varied in severity fromsuperficial abrasions to deep lacerations and oculartrauma. Other organ injuries were also noted in four pa-tients. One patient (patient 5) had “multiple trauma” witha hemopneumothorax, a fractured right clavicle, a closedleft tibia/fibula fracture, and bilateral cranial nerve 6 pal-sies. Not all children had an associated cervical spinalcord injury. Of the three surviving patients, two had aconcomitant spinal cord injury. At presentation to us, onehad a partial spinal cord lesion with preservation of sen-sation and mild spasticity in all extremities. However,she had good strength and motor function (patient 2).The other survivor with a spinal cord injury had a partial

cervical cord lesion with preservation of sensation butwas quadriplegic (patient 5).

Table 2 summarizes the findings from the initial lat-eral cervical spine radiographs. Two children were ini-tially seen in the emergency department at DUMC andthe diagnosis of traumatic AOD was made. The otherthree children were transferred to our institution afterstabilization/treatment was performed at other hospitals.In one of these patients, the diagnosis of traumatic AODwas made only after the evaluation in our emergencydepartment. The remaining two patients (patients 2 and5) were transferred to DUMC and directly admitted tothe pediatric rehabilitation hospital (Lenox Baker Chil-dren’s Hospital) for rehabilitation after a brain injury.During the admission physical examination, the diagno-sis of a cervical spinal cord injury was made. In both ofthese patients, it was the presence of spasticity (briskdeep tendon reflexes and clonus) in all extremities in thecontext of the lack of significant cognitive impairmentthat prompted us to evaluate for potential cervical spinalcord injury. The diagnosis of traumatic AOD was notmade initially by the lateral cervical plain film radio-graphs. Traumatic AOD was diagnosed by computed to-mography (CT) of the cervical spine (patient 2) (Fig. 4)and by magnetic resonance imaging of the craniocervicaljunction (patient 5).

We missed the diagnosis of traumatic AOD at the timeof transfer when the initial cervical spine radiographswere evaluated in these two children. Retrospective mea-surements of these radiographs revealed the presence oftraumatic AOD. For patient 2 the DB distance was 7 mm,with a ratio of Powers of 1.35. Patient 5 had a DB dis-tance of 2 mm, with a ratio of Powers of 1.11. Whenaveraging these measurements for all five patients, a DB

FIG. 2. Measurement for the ratio of Powers (BC:OA), where Brepresents the basion, C the posterior arch of the atlas, O theopisthion, and A the anterior arch of the atlas.

FIG. 1. Measurement of the distance from basion (B) to dens (D).

FIG. 3. Typical lateral cervical spine radiograph showing our cal-culations for the dens to basion distance (DB) and the ratio ofPowers (BC/OA).

K. KENTER ET AL.586

J Pediatr Orthop, Vol. 21, No. 5, 2001

20

distance of 9.8 mm and a ratio of Powers of 1.38 werecalculated.

The two children initially seen in the emergency de-partment at DUMC were immediately diagnosed withtraumatic AOD but died in the pediatric intensive careunit, one at 24 hours and one at 4 days. The other threechildren were placed in halo immobilization once trau-matic AOD was diagnosed. All of these children under-went posterior occiput-vertebral (C2 or C3) fusion. Ofthe two children with associated spinal cord injury, one(patient 2) has made a complete recovery. The otherchild (patient 5) remains quadriplegic with severe med-ullary injuries. At 10 year follow-up she has a tracheos-tomy and requires assisted ventilation at night. She doesnot have voluntary control of respiration, has palatal my-oclonus, has no functional use of her hands, and has onlyminimal antigravity movement of her iliopsoas, ham-string, and quadriceps musculature. A gastrostomy tubefeeds her. She is aphonic as a result of damage to thenucleus ambiguus resulting in vocal cord paralysis, butcan communicate by mouthing words. She has graduatedfrom high school and attends a community college.

DISCUSSION

The ligamentous complex at the level of the superiorfacets of the atlas supplies the axial strength at the cra-niovertebral junction. This complex consists of the ante-rior and the posterior atlanto-occipital membranes andthe two lateral atlanto-occipital ligaments. The majorityof strength at the craniocervical junction is due to theatlanto-occipital ligamentous structure (25). The atlasring is seated within a ligamentous complex joining theocciput to the axis. This complex consists of the tectorialmembrane, the anterior longitudinal ligament, and thenuchal ligament. The apical dental ligament and thepaired alar ligaments also play a role in stability. Werne(24) showed in a classic study that forward flexion of thecraniovertebral articulation is limited by the bony contactof the dens on the anterior aspect of the foramen mag-

num. He further found that hyperextension is limited bythe tectorial membrane, whereas lateral flexion and ro-tation are limited by the alar ligaments and the occipito-cervical capsules.

The mechanism for traumatic AOD remains incom-pletely understood. Some investigators believe that ex-treme hyperextension of the cranium on the spine leadsto rupture of the tectorial membrane, causing traumaticAOD (8,21). Others have suggested that a component oflateral flexion is necessary to disrupt the alar ligaments(12,18). Finally, Eismont and Bohlman (9) believe thatall of the ligamentous structures must be completely dis-rupted between the atlantoaxial articulation for traumaticAOD to occur. All five children in our series were vic-tims of high-energy impacts with the possibility of ex-treme deceleration forces; however, the actual mecha-nism for AOD was not studied.

Traumatic AOD may be relatively more common inchildren than in adults (3,19). The craniovertebral junc-tion in children may be less stable than in adults becausethe occipital condyles are relatively smaller and the ar-ticulation between the cranium and the atlas is morehorizontal (3,14). As the child grows, the articulationbecomes more deeply seated in the superior facets of theatlas and therefore more stable.

Given the intrinsic stability of the craniocervical junc-tion, a high-energy impact is necessary to cause trau-matic AOD. This impact results frequently in brain andfacial injuries as well (7). Of the five patients describedhere, four had concomitant closed head injuries and fourhad facial injuries. One child with traumatic AOD hadrelatively minor facial injuries and neck pain but no evi-dence of brain or spinal cord injuries. This shows thattraumatic AOD is not always associated with severebrain, spinal cord, or facial injuries.

Two of the three children transferred to DUMC hadcervical spinal cord injuries. Both of these patients had

TABLE 2. Radiologic findings from initial lateral cervicalspine radiographs

Case no. Retropharyngeal swelling Fracture AOD diagnosed

1 No No Yes2 No No No3 Yes No Yes4 Yes No Yes5 No No No

AOD, atlanto-occipital dislocation.

FIG. 4. Coronal reconstruction from a computed tomographyscan showing the diagnosis of atlanto-occipital dislocation. Ar-rows point to the occipital condyles. Note the lateral translation ofthe occiput with respect to the odontoid.

TABLE 1. Clinical features of five pediatric cases of traumatic AOD from 1985 to 1995

Case no. Gender Age (yrs) Cause CHI Facial injuries Other organ injury Spinal cord injury Outcome

1 M 7 MVC Yes No No Yes Died 24 h2 F 9 MVC Yes Yes Yes Yes Survived3 F 9 Pedestrian Yes Yes Yes Yes Died 4 days4 M 12 Pedestrian No Yes Yes No Survived5 F 9 MVC Yes Yes Yes Yes Survived

CHI, closed head injury; MVC, motor vehicle crash.

TRAUMATIC ATLANTO-OCCIPITAL DISLOCATION 587


21

spasticity in a widespread anatomical distribution andminor to no cognitive impairment. For a brain injury tohave caused spasticity in such a distribution, it is likelythat the patients would also have had a more prolongedperiod of coma and severe cognitive impairment. Ab-sence of these features localized injury to the cervicalspinal cord.

Sneed and Stover (22) reported undiagnosed spinalcord injuries in four brain-injured children. The physicalfindings that can lead to the diagnosis of cervical spinalcord injury in patients with a concomitant brain injuryare summarized in Table 3. It is thought that quadriplegiaand quadriparesis are common in survivors of traumaticAOD (11,18,19). However, only one of the three chil-dren who survived in our series was quadriparetic. Pangand Wilberger (18) described some of their patients ashaving a Brown-Sequard pattern of hemiplegia. Otherauthors reported that some patients had no neurologicdeficit (9,12). One of our patients with a spinal cordinjury eventually recovered completely, with no residualdeficit.

Cranial nerve injuries are also frequently reported incases of traumatic AOD (8,11,12,16,21). Upper cranialnerve palsies may be related to concomitant brain injury(especially injury to cranial nerve 6, as seen in our pa-tient 5) (8,12,19–21). Lower cranial nerve and medullaryinjuries may result from direct trauma to the brain stem(14,19,20,23), from compression of the brain stem by anexpanding extradural hematoma (13,19,20,23), or fromtraction on nerve roots and vessels resulting from dis-placement of the occiput from the atlas (11). Vascularcompression, vasospasm, tears of the vertebral-basilararterial system, or artery thrombosis can also lead toinjury to the brain stem (8,12,20). Pang and Wilberger(18) emphasized the difficulty in diagnosing medullaryand complete high cervical spinal cord injuries in pa-tients with traumatic AOD. Respiratory distress, apnea,and neurogenic shock also occur in patients with severebrain injuries alone who have not suffered traumaticAOD. Death from traumatic AOD is probably a conse-quence of both a complete high cervical spinal cord in-jury causing diaphragmatic paralysis and a medullary

injury resulting in central respiratory apnea. The radio-logic findings of traumatic AOD on the lateral cervicalspine radiographs may be subtle and go unrecognized, aswas the case for the three children initially undiagnosedin this series. Retropharyngeal soft tissue swelling waspresent in only two of the five children, supporting thecontention that this finding lacks sensitivity (15). Retro-pharyngeal swelling has been reported by other authorsto occur in 80% of patients with traumatic AOD (4).

Wholey et al. (26) analyzed 600 lateral cervical radio-graphs to determine the normal distance between the tipof the dens and the basion (DB distance). The normal DBdistance was !5 mm in adults and !10 mm in children.Kaufman et al. (14) measured the distance between theoccipital condyle and the superior facet of the atlas innormal children who were 1 to 15 years of age and foundthat it was <5 mm. Both of these methods have beencriticized as being unreliable, despite the ease of mea-surement, because of the variations in radiographic mag-nification, the variations in the amount of neck flexionand extension or with head rotation, the variations in thethickness of the articular cartilage covering the occipitalcondyles, and the difficulty in identifying the bony land-marks (8,13,14,18,20). Powers et al. (20) attempted toincrease the accuracy of radiologic diagnosis by devel-oping a method of interpretation unaffected by magnifi-cation. The ratio of Powers is the distance from the ba-sion to the posterior arch on the atlas divided by thedistance from the opisthion to the anterior arch of theatlas (see Fig. 2). This ratio has a mean value of 0.77 ina normal population. A ratio >1.0 is considered diagnos-tic for traumatic AOD. This ratio misses the rare patientwith pure longitudinal distraction or posterior atlanto-occipital dislocation (14,27). In our series, the ratio ofPowers was more accurate than the DB distance for thediagnosis of traumatic AOD. This was present only in thetwo children whose diagnosis of traumatic AOD wasmissed on the initial lateral cervical plain film radio-graphs.

Bulas et al. (4) found that in all 11 cases of traumaticAOD, a DB distance >14 mm was measured. In theirreport, none of the 110 pediatric trauma patients withoutAOD had this great a DB distance. In contrast, the ratioof Powers was >1.0 in only 6 of their 11 patients withtraumatic AOD. Combining their results with the resultspresented here, one must conclude that both the DB dis-tance and the ratio of Powers should be measured on thelateral cervical spine radiograph in patients suspected ofhaving traumatic AOD.

Some authors have advocated the use of CT or mag-netic resonance imaging instead of lateral cervical spineradiographs to make the diagnosis of traumatic AOD(10,13,16). CT of the cervical spine has some advantagesover cervical spine radiographs, including ease of patientpositioning for multiple planes, greater anatomical bonydetail, and the ability to obtain the same anatomical im-ages over time, thus enabling a better evaluation of treat-ment, if needed. Although the CT findings confirmed thediagnosis of traumatic AOD in one of our patients, ret-rospective measurements from the initial lateral cervical

TABLE 3. Signs of cervical spinal cord injury in presenceof brain injury

Severe spinal cord injurySpinal shock (early)Muscle flaccidity (early)Absent deep tendon and sacral reflexes (early)PoikilothermiaSevere spasticity (late)Urinary retention (late)Priaprism (late)Autonomic dysreflexia (late)

Less severe spinal cord injuryNeck painCervical muscle spasmsDiffuse tingling (“sprangles”)Spasticity > cognitive impairmentDermatomal sensory loss

Modified from Sneed and Stover (22), with permission.

K. KENTER ET AL.588


22

spine radiographs did or could have made the diagnosis.Nonetheless, we recommend that CT sagittal images ofthe cervical spine to the level of the second vertebra (C2)be performed. This could be performed when the brainCT scan is obtained in patients with closed head injuries.We are aware, however, that CT images of the cervicalspine alone failed to show nontraumatic AOD in two offour patients described by Cohn et al. (5). Thus, CTimaging alone may not be sufficient to rule out traumaticAOD.

In summary, from our experience with these five casesof traumatic AOD in children, the following conclusionscan be drawn. Neck pain and/or signs and symptoms ofspinal cord injury in children who have suffered signifi-cant trauma should prompt an investigation of cervicalspine injury, including traumatic AOD. Not all pediatricpatients with traumatic AOD have concomitant severeclosed head or facial injuries. The diagnosis of traumaticAOD by lateral cervical spine radiographs can be verydifficult. When the diagnosis of traumatic AOD is sus-pected, the ratio of Powers and the DB distance shouldbe measured. CT evaluation of the brain after significanthead or craniocervical trauma should routinely includeimages to C2 as a screening assessment for traumaticAOD. Prompt recognition and treatment of traumaticAOD may prevent severe spinal cord injuries in somepediatric patients.

REFERENCES1. Alker GJ, Jr, Oh YS, Leslie EV. High cervical spine and cranio-

cervical junction injuries in fatal traffic accidents. Orthop ClinNorth Am 1978;9:1003–10.

2. Blackwood NJ. Alto-occipital dislocation. A case of fracture of theatlas and axis, and forward dislocation of the occiput on the spinalcolumn, life being maintained for thirty-four hours and forty min-utes by artificial respiration, during which a laminectomy wasperformed upon the third cervical vertebra. Ann Surg 1908;47:654–8.

3. Bucholz RW, Burkhead WZ. The pathological anatomy of fatalatlanto-occipital dislocations. J Bone Joint Surg [Am] 1979;61:148–250.

4. Bulas DI, Fitz CR, Johnson DL. Traumatic atlanto-occipital dis-location in children. Radiology 1993;188:155–8.

5. Cohn A, Hirsch M, Katz M, et al. Traumatic atlanto-occipitaldislocation in children: review and report of five cases. PediatrEmerg Care 1991;7:24–7.

6. Davis D, Bohlman H, Walker AE, et al. The pathological findingsin fatal craniospinal injuries. J Neurosurg 1971;34:603–13.

7. Dibenedetto T, Lee CK. Traumatic atlanto-occipital instability. Acase report with follow-up and a new diagnostic technique. Spine1990;15:595–7.

8. Dublin AB, Marks WM, Wienstock D, et al. Traumatic dislocationof the atlanto-occipital articulation (AOA) with short-term sur-vival. With a radiographic method of measuring the AOA. J Neu-rosurg 1980;52:541–6.

9. Eismont FJ, Bohlman HH. Posterior atlanto-occipital dislocationwith fractures of the atlas and odontoid process. J Bone Joint Surg[Am] 1978;60:397–9.

10. Farley FA, Graziano GP, Hensinger RN. Traumatic atlanto-occipital dislocation in a child. Spine 1992;17:1539–41.

11. Fruin AH, Pirotte TP. Traumatic atlantooccipital dislocation. Casereport. J Neurosurg 1977;46:663–6.

12. Gabrielsen TO, Maxwell JA. Traumatic atlanto-occipital disloca-tion; with case report of a patient who survived. AJR Am J Roent-genol 1966;97:624–9.

13. Gerlock AJ Jr, Mirfahkraee M, Benzel EC. Computed tomographyof traumatic atlanto-occipital dislocation. Neurosurgery 1983;13:316–9.

14. Kaufman RA, Carroll CD, Buncher CR. Atlantooccipital junction:standards for measurement in normal children. AJNR Am J Neu-roradiol 1987;8:995–9.

15. Lee C, Woodring JH, Goldstein SJ, et al. Evaluation of traumaticatlantooccipital dislocation. Am J Neuroradiol 1987;8:19–26.

16. Matava MJ, Whitesides TE, Davis PC. Traumatic atlanto-occipitaldislocation with survival. Spine 1993;18:1897–903.

17. Montane I, Eismont FJ, Green BA. Traumatic occipitoatlantal dis-location. Spine 1991;16:112–6.

18. Pang D, Wilberger JE Jr. Traumatic atlanto-occipital dislocationwith survival: case report and review. Neurosurgery 1980;7:503–8.

19. Papadopoulos SM, Dickman CA, Sonntag VKH, et al. Traumaticatlantooccipital dislocation with survival. Neurosurgery 1991;28:574–9.

20. Powers B, Miller MD, Kramer RS, et al. Traumatic anterior at-lanto-occipital dislocation. Neurosurgery 1979;4:12–7.

21. Rockswold GL, Seljeskog EL. Traumatic altantocranial dislocationwith survival. Minn Med 1979;62:151–4.

22. Sneed RC, Stover SL. Undiagnosed spinal cord injuries in brain-injured children. Am J Dis Child 1988;142:965–7.

23. Traynelis VC, Marano GD, Dunker RO, et al. Traumatic atlanto-occipital dislocation. Case report. J Neurosurg 1986;65:863–70.

24. Werne S. Studies in spontaneous atlas dislocation. Acta OrthopScand 1957;23(Suppl):1–150.

25. White AA III, Panjabi MM. The clinical biomechanics of the oc-cipito–atlantoaxial complex. Orthop Clin North Am 1978;9:867–78.

26. Wholey MH, Bruwer AJ, Baker HL Jr. The lateral roentgenogramof the neck (with comments on the atlanto-odontoid-basion rela-tionship). Radiology 1958;71:350–6.

27. Woodring JH, Selke AC Jr, Duff DE. Traumatic atlantooccipitaldislocation with survival. AJR Am J Roentgenol 1981;137:21–4.

TRAUMATIC ATLANTO-OCCIPITAL DISLOCATION 589


23

Journal of Pediatric SurgeryVOL 36, NO 8 AUGUST 2001

Cervical Spine Injuries in Children: A Review of 103 PatientsTreated Consecutively at a Level 1 Pediatric Trauma Center

By Rebeccah L. Brown, Margie A. Brunn, and Victor F. GarciaCincinnati, Ohio

Purpose: Cervical spine (C-spine) injuries occur infrequentlyin children but may be associated with significant disabilityand mortality. The purpose of this study was to review theexperience of a level 1 pediatric trauma center to determinethe epidemiology, risk factors, mechanisms, levels, types ofinjury, comorbid factors, and outcomes associated withthese potentially devastating injuries.

Methods: A retrospective analysis of 103 consecutive C-spine injuries treated at a level 1 pediatric trauma center overa 91⁄2-year period (January 1991 through August 2000) wasperformed.

Results: The mean age was 10.3 ! 5.2 years, and the male-to-female ratio was 1.6:1. The most common mechanism ofinjury was motor vehicle related (52%), followed by sportinginjuries (27%). Football injuries accounted for 29% of allsports-related injuries. Sixty-eight percent of all children sus-tained injuries to C1 to C4; 25% to C5 to C7; and 7% to both.Spinal cord injury without radiographic abnormality (SCI-WORA) occurred in 38%. Five patients had complete cordlesions involving the lower C-spine (C4 to C7); 4 of thesewere motor vehicle related, and all 4 patients died. IsolatedC-spine injuries occurred in 43%, whereas 38% had associ-ated closed head injuries (CHI). The overall mortality ratewas 18.5%, most commonly motor vehicle related (95%),occurring in younger children (mean and median age 5years) and associated with upper C-spine injuries (74%) andCHI (89%). C1 dislocations occurred in younger children(mean age, 6.6 years), most often as a result of motor vehi-cle–related trauma (especially pedestrians) and were associ-ated with the highest injury severity score (ISS), longest

length of stay (LOS), most CHIs, and the highest mortalityrate (50%). C-spine fractures with or without SCI occurredmost commonly as a result of falls and dives. Sportinginjuries occurred almost exclusively in adolescent boys(mean age, 13.8 years) and were isolated injuries associatedwith a relatively low ISS and shorter LOS. Interestingly, 75%of sporting injuries showed SCIWORA, and all infants suffer-ing from child abuse had SCIWORA.

Conclusions: Mechanisms of injury are age related, withyounger children sustaining C-spine injuries as a result ofmotor vehicle–related trauma and older adolescents com-monly injured during sporting activities. C-spine injuries inchildren most commonly involve the upper C-spine, butcomplete lesions of the cord are associated more frequentlywith lower C-spine injuries. The type of C-spine injury isrelated to the mechanism of injury: SCIWORA is associatedwith sporting activities and child abuse, C-spine dislocationsmost commonly result from motor vehicle–related trauma(especially among pedestrians), and C-spine fractures occurmost commonly as a result of falls and dives. Predictors ofmortality include younger age, motor vehicle–related mech-anism, C1 dislocations, high ISS greater than 25, and asso-ciated CHI. A high index of suspicion for SCIWORA is essen-tial when evaluating adolescents with neck traumaassociated with sporting injuries or victims of child abuse.J Pediatr Surg 36:1107-1114. Copyright © 2001 by W.B.Saunders Company.

INDEX WORDS: Cervical spine injuries, pediatric, trauma,spinal cord injury without radiographic abnormality.

From the Children’s Hospital Medical Center, Division of Trauma Services, Cincinnati, OH.Presented at the 2000 Annual Meeting of the Section on Surgery of the American Academy of Pediatrics, Chicago, Illinois, October 28-November

1, 2000.Address reprint requests to Rebeccah L. Brown, MD, Assistant Professor of Clinical Surgery and Pediatrics, Associate Director of Trauma

Services, Children’s Hospital Medical Center, Division of Pediatric Surgery, OSB-3, 3333 Burnet Ave, Cincinnati, OH 45229-3039.Copyright © 2001 by W.B. Saunders Company0022-3468/01/3608-0001$35.00/0doi:10.1053/jpsu.2001.25665

1107Journal of Pediatric Surgery, Vol 36, No 8 (August), 2001: pp 1107-1114

24

ABOUT 1% to 2% of all pediatric patients requiringhospital admission for traumatic injuries will have

injuries to the cervical spine (C-spine).1,2 Although theincidence is relatively low, C-spine injuries in childrencan have devastating consequences with enormous emo-tional, social, and economic impact.3,4 A thorough un-derstanding of C-spine injuries is essential so that theseinjuries are not overlooked. The purpose of this studywas to review our experience with C-spine injuries inchildren admitted to a level 1 pediatric trauma center todetermine the epidemiology, risk factors, mechanisms ofinjury, level and types of injury, comorbid factors, pre-dictors of outcome and mortality, and to identify possiblestrategies for prevention.

MATERIALS AND METHODSWe queried our trauma registry for all children admitted to Chil-

dren’s Hospital Medical Center, Cincinnati, OH, an accredited level 1trauma center, between January 1991 and August 2000 with cervicalspine injuries (ICD-9 codes 805.0 to 805.1, closed or open C-spinefractures without spinal cord injury; 806.0 to 806.19, closed or openC-spine injuries with spinal cord injury; 839.0 to 839.1, closed or openC-spine dislocations; and 952.0 to 952.09, cervical spinal cord injurywithout evidence of spinal bony injury or SCIWORA). Trauma Reg-istry data were corroborated by chart review. This study was approvedby the Children’s Hospital Medical Center Institutional Review Board.Over the 91⁄2-year period, there were 4,619 trauma admissions, and,

of these, 103 (2.2%) sustained cervical spine (C-spine) injuries. Dataextracted from the trauma registry and chart review included age, race,sex, injury environment, mechanism of injury, usage of injury preven-tion devices (ie, safety belts, child restraint systems), initial emergencyroom Glasgow Coma Score (GCS), injury severity score (ISS), level ofC-spine injury, presence of associated spinal cord injury (SCI), pres-ence of closed head injuries (CHI) or other traumatic injuries, length ofhospital stay (LOS), need for operative intervention, disposition, andmortality rate. Because of known differences in the anatomy, biome-chanics, and injury patterns in children !8 years of age, patients weredivided into 2 groups for comparison purposes: younger children (0 to8 years) and older children and adolescents (9 to 19 years).

RESULTS

DemographicsThere were 103 children admitted to Children’s Hos-

pital Medical Center, Cincinnati, OH with C-spine inju-ries during the 91⁄2-year study period. The admission ratewas relatively steady at about 10 to 12 admissions peryear. The mean age was 10.3 ! 5.2 years (range, 2months to 19 years). About two thirds were more than 8years of age. The age distribution (Fig 1) was bimodalwith a peak incidence in the 13- to 15-year age group anda smaller peak around 5 years of age. Overall, injuredboys outnumbered girls 1.6:1. For motor vehicle-relatedmechanisms, the male-to-female ratio was almost equal(1.2:1); however, for sports-related injuries, boys pre-dominated (3.5:1). Eighty percent of children werewhite, 16% were African-American, and 4% were ofother racial origins.

Mechanism of InjuryAmong all age groups, the predominant mechanism of

C-spine injury (Table 1) was motor vehicle related(52%): motor vehicle crashes (MVCs; 31%), pedestrianversus motor vehicle (16%), and bicycle versus motorvehicle (6%). Sports-related activities accounted for 27%of all injuries, followed by falls (15%), child abuse (3%),and unknown (3%).Motor vehicle–related injuries. Fifty-four children

were injured in motor vehicle–related incidents: 37 (69%)sustained upper C-spine injuries (C1-C4), 11 (20%) hadlower C-spine injuries (C5-C7), and 6 (11%) had multi-ple C-spine fractures or dislocations. Of the 32 passen-Fig 1. Age distribution of children with cervical spine injuries.

Table 1. Mechanisms of Injury in ChildrenWith Cervical Spine Injuries

Mechanisms of Injury

No. of Cervical Spine Injuries (n " 103)

Age

0 to 8 yr 9 to 19 yr All

Motor vehicle related 25 29 54 (52%)Passenger 17 15 32Pedestrian 7 9 16Bicyclist 1 5 6

Sports related 28 28 (27%)Football 8 8Diving 5 5Soccer 3 3Wrestling 3 3Trampoline 2 2Basketball 2 2Sledding 2 2Baseball 1 1Hockey 1 1Water skiing 1 1

Falls 5 10 15 (15%)Child abuse 3 0 3 (3%)Other 2 1 3 (3%)Totals 35 68 103

1108 BROWN, BRUNN, AND GARCIA

25

gers or drivers injured in MVCs (mean age, 8.5 years;range, 6 months to 19 years), 18 had upper C-spineinjuries, 8 had lower C-spine injuries, and 6 had multipleC-spine fractures or dislocations. Of the 16 pedestriansstruck by motor vehicles (mean age, 9 years; range, 2 to15 years), 14 had upper C-spine injuries, including 6with C1 dislocations. Furthermore, 71% of younger pe-destrians had C1 dislocations. Three pedestrians hadcomplete spinal cord lesions: C4 level complete SCI-WORA in 1, C4 fracture with complete spinal cordlesion in 1, and C5-C7 fracture with complete spinal cordlesion in 1. One adolescent pedestrian sustained a fatalopen C1 to C4 fracture with multiple associated injuries.Of the 6 bicyclists (mean age, 13.8 years; range, 5 to 16years), all but one sustained upper C-spine injuries.Upper C-spine injuries predominated in both the

younger and older age groups involved in motor vehicle–related trauma (72% and 62%, respectively). Cervicalspine dislocations (mostly C1) occurred most commonlyin younger children, especially pedestrians, whereas C-spine fractures with and without SCI predominated inolder children and adolescents.Children sustaining C-spine injuries from motor vehi-

cle–related trauma had the highest mean ISS (25) and thehighest mortality rate (33%). Thirty-one of 54 had asso-ciated CHIs, with a mean GCS of 9.3, and almost half(46%) had an initial GCS of 8 or less on arrival to theemergency room. The mean LOS was 14 days. As mightbe expected, younger children had a higher mean ISS(29.8 v 20.5), more CHIs (72% v 45%), lower mean GCS(6.7 v 11.4), longer LOS (23 days v 6 days), and highermortality rate (48% v 21%) than older children andadolescents.Sports-related injuries. Sports-related C-spine inju-

ries occurred exclusively in the older age group, with amean age of 13.8 years. Almost all (96%) were white,and 78% were boys. Of the 28 sports-related injuries,almost one-third (29%) occurred while playing football.Diving (n ! 5), soccer (n ! 3), and wrestling (n ! 3)were the next most common sports-related mechanismsof injury followed by basketball (n! 2), trampoline (n!2), sledding (n ! 2), baseball (n ! 1), hockey (n ! 1),and water-skiing (n ! 1).Seventy-five percent of sports-related injuries in-

volved the upper C-spine (C1 to C4). The most strikingfinding among children with sports-related C-spine inju-ries was the overwhelming predominance of SCIWORAassociated with this mechanism of injury. Twenty-one of28 children had SCIWORA: 20 at the C1 to C4 level andone at the C5-C7 level. Seven of 8 football players hadSCIWORA, and all soccer, wrestling, and basketballinjuries exhibited SCWIORA. Fortunately, most childrenand adolescents with sports-related SCIWORA had tran-sient neurologic symptoms and recovered without ad-

verse sequelae, as reflected by their short mean LOS (2.5days).Divers sustained the most severe injuries of all sports-

related trauma: C1 to C4 SCIWORA with central cordsyndrome (n ! 1), C5 to C7 fracture with complete cordtransection (n ! 1), C5 to C7 fracture with anterior cordsyndrome (n ! 1), and C5 to C7 fracture without SCI(n ! 2). Interestingly, 80% involved the lower ratherthan the upper C-spine. Because of the presence ofsignificant neurologic deficits (60%), the mean LOS forthose with diving injuries was much longer (11.8 days)than for other types of sporting injuries.Overall, those with sports-related C-spine injuries

fared well. The majority (86%) had isolated injuries,with a mean ISS of 10, and the mean LOS for all patientswith C-spine injuries caused by sporting activities wasonly 4.5 days. There was no mortality associated withthis mechanism of injury.Falls. The mean age for children sustaining C-spine

injuries caused by falls was 11 years, and 10 of the 15were more than 8 years of age. About one third (27%)fell out of windows, 13% fell down stairs, and theremaining falls were quite varied (ie, from a tree, furni-ture, hot tub, playground equipment, farming equipment,being pushed).Two-thirds of C-spine injuries caused by falls in-

volved the upper C-spine. However, age-related differ-ences were observed, with 80% of injuries involving theupper C-spine in younger children versus only 50% inolder children. Almost half had isolated fractures withoutSCI. Four cases of SCIWORA were attributed to falls: 3at the C1 to C4 level, and 1 at the C5 to C7 level. C-spineinjuries caused by falls tended to be isolated, althoughyounger children who fell were 6 times more likely tohave an associated minor CHI. As with sport-relatedinjuries, the mean ISS and LOS for falls were relativelylow (8 and 7.3 days, respectively), and there were nodeaths.Child abuse. Child abuse occurred in 3 infants under

the age of 1 year. Two were white; 1 was African-American. Interestingly, all 3 patients had SCIWORA: 2at both the C1 to C4 and C5 to C7 levels (1 of these hadcentral cord syndrome) and 1 at the C5 to C7 level withan incomplete spinal cord lesion and posterior cord syn-drome. All had severe associated injuries: 1 had a CHI,multiple fractures of the radius, and a poison ingestion; 1had a severe associated CHI; and 1 had a pulmonarycontusion, pneumothorax, liver laceration, and fracturesof the ribs, scapula, humerus, femur, and thoracic spine.This latter patient died of his extensive injuries, whereasthe other 2 survived with significant deficits.Similar to victims of motor vehicle–related trauma,

these children had a high mean ISS (25), long mean LOS(15.7 days), and a high mortality rate (33%). The 2

1109CERVICAL SPINE INJURIES

26

survivors were placed in foster homes after prolongedstays in the rehabilitation unit.

Level and Types of Cervical Spine InjuryOf 103 children with cervical spine injuries, the ma-

jority (68%) sustained injuries to the upper C-spine (C1to C4). The distribution of these injuries is depicted inTable 2. About half of upper cervical spine injuries werecaused by C1 to C4 level SCIWORA and occurredpredominantly in the older age group, mostly sportsrelated. Sixteen (22%) of those with upper C-spine inju-ries had dislocations, the most common of which was aC1 dislocation (94%). Upper C-spine fractures withoutspinal cord injury (SCI) occurred in 14 patients (13%).The most commonly fractured upper cervical vertebraewas C2 (6 of 14). The remaining 8 patients (11%) withupper C-spine injuries had C-spine fractures with SCI.Twenty-five percent of patients had lower C-spine

injuries (C5 to C7), half of which were fractures withoutspinal cord injury. The most commonly fractured lowercervical vertebrae was C7 (85%). Four patients sustainedfractures at the C5 to C7 level with SCI. There were only2 dislocations of the lower C-spine, both of whichoccurred from motor vehicle-related trauma. There were7 cases of C5 to C7 level SCIWORA, almost evenlydistributed between the younger and older age groups.Of younger patients, 68% had injuries involving the

upper C-spine (C1 to C4), 24% the lower C-spine (C5 toC7), and 8% multiple sites. Of the older patients, asimilar distribution of upper, lower, and multiple C-spineinjuries was observed (69%, 25%, and 6%, respectively).

C-spine fractures without SCI. C-spine fractureswithout SCI accounted for 30% of all injuries in thisstudy. Almost 90% occurred in the older children andadolescents, and most were caused by falls or motorvehicle–related trauma. Although about two thirds hadassociated injuries, including 32% with associated CHIs,the mean ISS and LOS for patients with C-spine frac-tures without SCI were relatively low (14.4 and 6 days,respectively). The overall mortality rate associated withthese injuries was 10%, but approached 30% in theyounger age groups.C-spine fractures with SCI. C-spine fractures with

SCI accounted for 13% of all injuries in this study. LikeC-spine fractures without SCI, these injuries occurredmost commonly in the older age group (67%) and mostwere caused by falls and motor vehicle–related trauma.Similarly, 54% had associated injuries, including 38%with CHIs. The major difference between the 2 injurytypes, however, was an increased mean LOS (14 days)and mortality rate (31%) among those with SCI.C-spine dislocations. Although C-spine dislocations

accounted for only 19% of all injuries in this study, theywere the most severe injuries of all. C-spine dislocationsoccurred predominantly (70%) in the younger age group(mean age, 6.6 years), and 80% were motor vehiclerelated, mostly pedestrian injuries. Interestingly, 40% ofpedestrians struck by motor vehicles sustained C1 dislo-cations. The upper C-spine (C1 to C4) was most com-monly involved, with C1 dislocations being most com-mon. Eighty percent of children had associated injuries(93% in the younger age group), including 65% withassociated severe CHIs (79% in the younger age group).Children with C-spine dislocations had the highest meanISS (26), longest mean LOS (26 days), lowest mean GCS(8), and the highest mortality rate (40%). For the youngerage group, the mortality rate approached 50%.SCIWORA. SCIWORA occurred in 38% of patients

with a mean age of 10.8 years. The majority were white(79%) and boys (72%). Over 80% of cases of SCIWORAinvolved the upper C-spine (C1 to C4). The mean ISSwas 14.3 (range, 9 to 42), and mean LOS was 4.6 days(range, 2 to 31 days). The mean ISS and LOS for motorvehicle–related SCIWORA was much higher than thatfor sports or falls. Sports-related activities were associ-ated with 21 of the 39 (54%) cases of SCIWORA in thisstudy. Moreover, SCIWORA accounted for 21 of 28(75%) sports-related C-spine injuries. SCIWORA tendedto be an isolated injury in 62% of cases. About one-thirdof patients had associated CHIs, usually minor. Themortality rate associated with SCIWORA was 8%. Ofthe 3 patients who died, 2 died of motor vehicle-relatedtrauma (1 passenger and 1 pedestrian), and the other wasa victim of child abuse. All of these had multiple other

Table 2. Level of Cervical Spine Injury

Level of C-Spine Injury

No. of C-Spine Injuries (n ! 105)*

Age

0 to 8 yrs 9 to 19 yrs All

Upper C-spine (68%)C1-C4 fracture 3 11 14C1-C4 fracture with SCI 4 4 8C1-C4 dislocation 10 6 16C1-C4 level SCIWORA 8 26 34Total upper C-spine 25 (24%) 47 (45%) 72 (68%)

Lower C-spine (25%)C5-C7 fracture 3 10 13C5-C7 fracture with SCI 0 4 4C5-C7 dislocation 2 0 2C5-C7 level SCIWORA 4 3 7Total lower C-spine 9 (9%) 17 (16%) 26 (25%)

Multiple C-spine (7%)Multiple C-spine fractures 1 3 4Multiple C-spine dislocations 2 1 3Total multiple C-spine 3 (3%) 4 (4%) 7 (7%)

Total all cervical spine injuries 37 68 105

*Two patients had both C1 to C4 and C5 to C7 SCIWORA, so totalnumber of injuries is 105 in 103 patients.


27

associated injuries, including 2 of the 3 with severe CHIsthat contributed significantly to their death.All children with SCIWORA (n ! 39) presented with

some type of neurologic deficit ranging from minor ortransient paresthesias or paresis to complete or perma-nent quadriplegia. All patients had initial C-spine radio-graphs with at least 2 views that were read as normal.Twenty-seven patients (69%) had further imaging of theC-spine with either magnetic resonance imaging (MRI)(n ! 25) or computed tomography (CT) (n ! 2); 8patients had no further imaging, and in the remaining 4patients, the radiographic records were not available forreview. The most common finding on MRI in patientswith SCIWORA was cord edema.Complete cord lesions. Five patients had complete

lesions of the spinal cord: 3 at the C4 level and 2 at theC5 to C7 level. One of these was a complete C4SCIWORA. Four of these were motor vehicle related—1passenger and 3 pedestrians. All of these patients hadassociated CHIs and other serious traumatic injuries, andall 4 died. The fifth patient sustained an isolated C5 toC7 fracture while diving and survived as a quadriplegic.The mortality rate associated with complete cord lesionswas 80%.

Associated InjuriesForty-four of 103 children had isolated C-spine inju-

ries, the majority of which were caused by sports-relatedtrauma (86%) or falls (62%). Fifty-nine patients hadother associated injuries, including CHIs in 39 children.Almost 90% of children who died had a concurrent CHI.Over 80% of children with motor vehicle–related

C-spine injuries had other associated injuries, with CHIsoccurring in 57%. The mean GCS on arrival to theemergency room for all patients with motor vehicle–related C-spine trauma and associated CHIs was 5.6(range, 3 to 15). The overall mortality rate for motorvehicle–related C-spine trauma with associated CHI was55%. Age-related differences were observed. Comparedwith older children, younger children were more likely tohave CHIs (72% v 45%), had lower mean GCS (4.7 v 7),and had higher mortality rate (61% v 46%).

MortalityOverall, 19 of 103 children with C-spine injuries died,

for a case fatality rate of 18.5%. The mean age fornonsurvivors was 5.3 years (range, 2 months to 15 years;median age 5 years). The mean ISS and LOS was 41.4(range, 25 to 75) and 4.2 days (range, 1 to 17 days;median, 2 days), respectively. All deaths but one wereattributable to motor vehicle–related trauma: 8 passen-gers, 8 pedestrians, and 2 bicyclists. The highest mortal-ity rate was for pedestrians (50%). Mortality rate was

25% for passengers in MVCs and 33% for bicyclists.There was 1 death caused by child abuse.Fourteen of the 19 children who died had upper

C-spine injuries; 5 had lower C-spine injuries. The pre-dominant type of injury was a C1 dislocation that oc-curred in 7 of the 19 nonsurvivors. Half of pedestrianfatalities involved C1 dislocations. Four patients hadcomplete cord lesions: 3 at the C4 level and 1 at the C5to C7 level. One patient had an open C1 to C4 fracturewith SCI. All patients had severe associated traumaticinjuries, including 17 of 19 (89%) with CHIs and a meanGCS of 3.5 (range, 3 to 9). Of the 2 without associatedCHIs, 1 had severe thoracic and cardiac injuries and theother had severe thoracic and abdominal injuries as wellas multiple fractures.

Operative InterventionNineteen of 103 children (18%) with C-spine injuries

required operative intervention. Indications for operativeintervention by either pediatric neurosurgery or orthope-dic surgery included cervical instability or need fordecompression of an incomplete SCI. Nine patients un-derwent anterior or posterior spinal fusion, 2 underwentplate fixation, and 8 had cervical halos placed. The mostcommon indications for operative intervention were un-stable C-spine fractures (n ! 8), followed by C-spinefractures with incomplete SCIs (n ! 5) and C-spinedislocations (n ! 5). Only one child with a C1 to C4SCIWORA required operative intervention—a halo wasplaced. The remainder of children with C-spine injurieswere treated with cervical immobilization in a rigidcollar or brace. Almost 60% of serious C-spine injuriesrequiring operative intervention were caused by motorvehicle–related trauma.

Use of Injury Prevention DevicesOf 32 motor vehicle passengers, only 6 (19%) were

reported to be appropriately restrained, whereas the re-maining 26 patients (81%) were either unrestrained orinappropriately restrained. Only 2 of the 8 children whodied in MVCs were restrained appropriately.In the younger children, 13 of 17 (76%) were either

unrestrained or inappropriately restrained, and the mor-tality rate was 38%. In the older children, 13 of 15 (87%)were totally unrestrained. The one patient in the olderage group who died was not restrained.Of the 6 bicyclists who sustained C-spine injuries,

none were wearing helmets. Both of the children whodied had severe associated CHIs.

DispositionOf the 103 children admitted to the hospital with

C-spine injuries, 68 were discharged home after a meanLOS of 10.9 days (range, 1 to 227 days); 16 were


28

transferred to our pediatric rehabilitation center for fur-ther therapy after a mean LOS of 16.5 days (range, 4 to47 days); and 19 died during their hospitalization.

DISCUSSION

The current review of 103 children admitted withcervical spine injuries to a level 1 pediatric trauma centerover a 91⁄2-year period represents one of the largersingle-institution studies in the literature to date. Theincidence of C-spine injuries in our institution over thestudy period was 2.2%, and this is consistent with thereported incidence of 0.65% to 9.47%.5-9 The demo-graphics of our patient population were similar to thoseof most other series in the literature.5,6,8-15 Similar to thefindings of Orenstein et al,11 we also identified a bimodalage distribution for patients sustaining cervical spineinjuries, with a large peak at around 13 to 15 years anda smaller peak around age 5 years. As in most otherseries, boys were more commonly injured than girls,with a male-to-female ratio of about 1.6:1. Adolescentmales accounted for the majority of sports-related C-spine trauma, whereas in the younger children, where thepredominant mechanism of injury was motor vehiclerelated, the sexes were more evenly distributed. As inother studies, we divided our population into 2 agegroups—0 to 8 years and 9 to 19 years—based on theunique anatomic and biomechanical differences in thecervical spine between these age groups.6,11,16-19 Therewere definite age-related differences in mechanism ofinjury. Younger children were more likely to sustainC-spine injuries caused by motor vehicle–related inci-dents, whereas older children and adolescents were morelikely to incur C-spine injuries during sporting endeavors.As opposed to the widespread opinion that upper

C-spine injuries occur in younger children and lowerC-spine injuries in older children, we noted that for allage groups and all mechanisms of injury, upper C-spineinjuries prevailed. Interestingly, all 5 complete cordtransections occurred at or below C4.Three distinct patterns of C-spine injury were identi-

fied in this study. First, SCIWORA, the most commontype of injury, occurred most commonly in adolescentboys engaged in sporting activities and in all abusedinfants. Second, upper C-spine dislocations occurredmost commonly in younger children and were causedprimarily by motor vehicle–related trauma. C1 disloca-tions predominated, especially among pedestrians. Theywere associated with multiple other injuries, primarilyCHIs, and carried a high mortality rate. Finally, C-spinefractures with and without SCI occurred most commonlyin older children and adolescents in association with fallsand dives.SCIWORA occurred in 38% of injured children. The

incidence of SCIWORA among children ranges from 4%

to 67%3,8,9,11,13,19-28 and is reported to occur predomi-nantly in younger children involved in motor vehicle–related incidents. However, in this series, over half of thecases of SCIWORA occurred in older children and ado-lescents with sports–related trauma. SCIWORA ac-counted for 75% of sporting injuries. Most children withSCIWORA presented with transient neurologic findings.All 21 patients with sports-related SCIWORA presentedwith short-lived paresthesias or motor weakness. Fortu-nately, none progressed to complete or permanent neu-rologic deficits. However, 10 of 18 patients with non–sports-related SCIWORA presented with more severeand worrisome neurologic deficits, including one with acomplete C4 spinal cord lesion; 7 with incomplete spinalcord lesions with posterior cord syndromes; and 2 withcentral cord syndromes. Two of these patients died. Bothwere involved in MVCs and had multiple other associ-ated injuries, including CHI. According to Pang andPollack,21 the sole predictor of outcome in SCIWORA isneurologic status at presentation. The majority of ourpatients had minimal neurologic deficits at presentation,thus explaining the excellent outcomes we observed.The possibility of exacerbating a neurologic injury

dictates a high index of suspicion for C-spine injury,even if the initial plain radiographs are negative. Patientsand parents must be thoroughly questioned about tran-sient weakness, paresthesias, numbness, “shocklike” sen-sations, or focal clumsiness after the traumatic event.18 Ifany of these symptoms are present, the patient’s C-spinemust be immobilized with an appropriate collar anddynamic flexion-extension films or an MRI scan ob-tained to rule out ligamentous instability or cord edemaor hemorrhage. Further complicating management is thefinding of Pang et al20,21 that over 50% of children withSCIWORA had delayed onset of paralysis up to 4 daysafter injury. In retrospect, most recalled transient pares-thesias, numbness, or subjective paralysis immediatelyafter the event. This underscores the importance of ob-taining a detailed history with regard to the presence ofneurologic symptoms. Pang and Pollack21 also describethe phenomenon of recurrent SCIWORA. One of ourpatients was admitted twice, once in 1996 and again in1999, with C1 to C4 SCIWORA, the first time because ofa basketball injury and the second time because of asoccer injury. It is suggested that the C-spine in childrenwith recurrent SCIWORA is rendered unstable by theinciting event, and is thus more prone to additional, oftenmore severe, neurologic trauma with secondary insults.Notably, the only sports-related C-spine injuries that

significantly deviated from the SCIWORA type of pre-sentation were diving injuries, with 80% involving thelower rather than the upper C-spine, 80% associated withC-spine fractures, 60% with significant neurologic defi-cits (one with complete cord transection), and only one


29

case of SCIWORA. Accordingly, the ISS and LOS werehigher for this population. It is presumed that the differ-ences in the types of injuries sustained is attributable tothe fact that the type of impact resulting from divinginjuries is significantly different than that resulting fromorganized sports-related activities like football, basket-ball, wrestling, and soccer (ie, head-on impact with ahard, unyielding surface versus human-to-human push-and-shove type contact).As shared by others, the most common mechanism of

injury in our study was motor vehicle related.8,10,11,14Motor vehicle–related trauma was associated with thehighest ISS, LOS, and mortality rate, especially amongthe younger children, whose mortality rate approached50%. Pedestrians struck by motor vehicles had the high-est mortality rate (50%), followed by bicyclists (33%),then passengers in MVCs (25%). Interestingly, all deathsbut one (95%), were attributable to motor vehicle–relatedtrauma.There was a strong correlation between pediatric C-

spine injuries and CHIs, especially with motor vehicle–related trauma. Moreover, CHI was a strong predictorof outcome and mortality. Overall, for all mechanismsof C-spine injury, 38% of patients had an associatedCHI with a mortality rate of 49%. In younger children,66% had an associated CHI with a 52% mortality rate,and in older children and adolescents, 24% had anassociated CHI with a mortality rate of 38%. As might beexpected, motor vehicle–related trauma was the mecha-nism of injury most commonly associated with CHIs.This strong correlation between pediatric C-spine inju-ries and CHIs is supported by the literature, in whichthere is a reported incidence varying from 30% to60%.8,11,14,22,29,30In addition to associated CHI and motor vehicle–

related mechanism of injury, other predictors of mortal-ity in children with C-spine injuries identified in thisstudy included younger age, C1 dislocations, and ISSgreater than 25. All nonsurvivors had ISS greater than25. It is noteworthy that the mean LOS for nonsurvivorsin this study was only 4.6 days. In fact, the median staywas only 2 days. These children did not linger in theintensive care unit over a prolonged course, and thisfinding most likely is a reflection of overall high injuryseverity with multiple associated injuries, including dev-astating CHIs.Child abuse accounted for 3% of the C-spine inju-

ries in this study. Interestingly, all abuse patients hadSCIWORA. Cervical spinal cord injuries usually are notincluded in the classic description of shaken baby syn-drome, but perhaps should be. Hadley et al31 reportedepidural or subdural hematomas of the spinal cord at thecervicomedullary junction in 5 of 6 child abuse victims,

as well as ventral spinal cord contusions in the upperC-spine in 4 of the 6 patients. MRI findings in ourpatients included diffuse edema and hemorrhage of thecervical spinal cord in 2 patients and global hypoperfu-sion of the spinal cord in the other patient who subse-quently died. Rooks et al32 also described 2 cases ofC-spine injury associated with child abuse. Although notclassically associated with child abuse, C-spine injuriesdo occur and should not be overlooked. One must main-tain a high index of suspicion for cervical SCIWORA inthe child abuse victim with normal C-spine radiographsand a worrisome neurologic examination finding. MRIcan be invaluable in this situation. If this type of injury isidentified early, expeditious treatment with steroids po-tentially could prevent further progression of cord edemaand neurologic symptoms.Because of the devastating sequelae of a C-spine

injury on the life of a child, prevention is a key concern.C-spine injuries may be prevented through appropriatemechanism-specific prevention strategies. The literatureclearly documents the effectiveness of safety belts andchild safety seats in reducing cervical spine injuries inchildren involved in MVCs; however, such restraintsmust be used properly, and the restraint system must beappropriate for the weight and age of the child.33-37 It isquite concerning that over 75% of younger children inour series were either totally unrestrained or not using anappropriate restraint device. Equally disturbing is the factthat 87% of older children and adolescents in this serieswere not restrained. For younger children, education ofparents regarding the choice and proper use of an appro-priate safety seat for their infants and toddlers and thenecessity of using a booster seat until the child is 8 yearsold and 80 pounds to protect from serious spinal injuriesmust be a top priority. For older children and adoles-cents, routine use of safety belts and their proper place-ment should be encouraged in health and physical edu-cation classes, youth group activities, and during sportsphysicals when other risk-taking behaviors are discussed.Parents and other adults must realize the importance ofserving as role models for their children.The incidence of cervical spine injuries resulting from

sports-related activities may be reduced by attention tophysical conditioning and strengthening, especially ofthe neck muscles, by football players and other athletes.Proper tackling and checking techniques as well asavoidance of head-butting maneuvers should be stressed.In general, all head-first moves should be avoided, ie,spearing in football, sliding head first in baseball, orhitting the boards head first during hockey. Proper fit offootball and hockey helmets so as not to extend belowthe nape of the neck may also reduce the likelihood ofcervical spine injuries.38 Spotters should be mandatory


30

when performing new or difficult moves in gymnastics.The importance of safe diving practices among teenagersalso cannot be overemphasized.C-spine injuries in children are potentially devastating,

resulting in significant morbidity and mortality. It is

incumbent on caregivers to maintain a high index ofsuspicion so that these injuries are not overlooked. Pre-dictable patterns of injury have been identified, andstrategies for prevention should be developed and imple-mented with these patterns in mind.

REFERENCES1. Jaffe DM, Binns H, Radkowski MA, et al: Developing a clinical

algorithm for early management of cervical spinal injury in childtrauma victims. Ann Emerg Med 16:270-275, 19872. Rachesky I, Boyce T, Duncan B, et al: Clinical prediction of

cervical spine injuries in children. Am J Dis Child 141:199-202, 19873. Anderson JM, Schutt AH: Spinal injury in children: A report of

156 cases seen from 1950 through 1978. Mayo Clin Proc 55:499-504,19804. Bohn D, Armstrong D, Becker L, et al: Cervical spine injuries in

children. J Trauma 30:463-469, 19905. Nitecki S, Moir CR: Predictive factors of the outcome of trau-

matic cervical spine fracture in children. J Pediatr Surg 29:1409-1411,19946. Hill SA, Miller CA, Kosnik EJ, et al: Pediatric neck injuries: A

clinical study. J Neurosurg 60:700-706, 19847. Kewalramani LS, Kraus JF, Sterling HM: Acute spinal cord

lesions in a pediatric population: Epidemiological and clinical features.Paraplegia 18:206-219, 19808. Eleraky MA, Therodore N, Adams M, et al: Pediatric cervical

spine injuries: Report of 102 cases and review of the literature.J Neurosurg 92:12-17, 20009. Ruge JR, Sinson GP, McLone DG, et al: Pediatric spinal injury:

The very young. J Neurosurg 68:25-30, 198810. Dietrich AM, Ginn-Pease ME, Bartkowski HM, et al: Pediatric

cervical spine fractures: Predominantly subtle presentation. J PediatrSurg 26:995-1000, 199111. Orenstein JB, Klein BL, Gotschall CS, et al: Age and outcome

of pediatric cervical spine injury: 11-year experience. Pediatr EmergCare 10:132-137, 199412. Baker C, Kadish H, Schunk JE: Evaluation of pediatric cervical

spine injuries. Am J Emerg Med 17:230-234, 199913. Finch GD, Barnes MJ: Major cervical spine injuries in children

and adolescents. J Pediatr Orthop 18:811-814, 199814. Givens TG, Polley KA, Smith GF, et al: Pediatric cervical spinal

injury: A three year experience. J Trauma 41:310-314, 199615. Hubbard DD: Injuries of the spine in children and adolescents.

Clin Orthop 100:56-65, 197416. Fesmire FM, Luten RC: The pediatric cervical spine: Develop-

mental anatomy and clinical aspects. J Emerg Med 7:133-142, 198917. Stauffer ES, Mazur JM: Cervical spine injuries in children.

Pediatr Ann 11:502-511, 198218. Kriss VM, Kriss TC: SCIWORA (spinal cord injury without

radiographic abnormality) in infants and children. Clin Pediatr 35:119-124, 199619. Hadley MN, Zabramski JM, Browner CM, et al: Pediatric spinal

trauma. Review of 122 cases of spinal cord and vertebral columninjuries. J Neurosurg 68:18-24, 1988

20. Pang D, Wilberger JE Jr: Spinal cord injury without radio-graphic abnormalities in children. J Neurosurg 57:114-129, 198221. Pang D, Pollack IF: Spinal cord injury without radiographic

abnormality in children—the SCIWORA syndrome. J Trauma 29:654-664, 198922. Birney TJ, Hanley EN Jr: Traumatic cervical spine injuries in

childhood and adolescence. Spine 114:1277-1282, 198923. Melzak J: Paraplegia among children. Lancet 2:45-48, 196924. Burke DC: Traumatic spinal paralysis in children. Paraplegia

11:268-27625. Hachen HJ: Spinal cord injury in children and adolescents:

Diagnostic pitfalls and therapeutic considerations in the acute stage.Paraplegia 15:55-64, 197726. Choi JU, Hoffman HJ, Hendrick EB, et al: Traumatic infarction

of the spinal cord in children. J Neurosurg 65:608-610, 198627. Osenbach RK, Menezes AH: Pediatric spinal cord and vertebral

column injury. Neurosurgery 30:385-390, 199228. Dickman CA, Zabramski JM, Hadley MN, et al: Pediatric spinal

cord injury without radiographic abnormalities: Report of 26 cases andreview of the literature. J Spinal Disord 4:296-305, 199129. Henrys P, Lyne ED, Lifton C, et al: Clinical review of cervical

spine injuries in children. Clin Orthop 129:172-176, 197730. Michael DB, Guyot DR, Darmody WR: Coincidence of head

and cervical spine injury. J Neurotrauma 6:177-189, 198931. Hadley MN: The infant whiplash-shake injury syndrome: A

clinical and pathological study. Neurosurgery 24:536-539, 198932. Rooks VJ, Sisler C, Burton B: Cervical injury in child abuse:

Report of two cases. Pediatr Radiol 28:193-195, 199833. Selecting and Using the Most Appropriate Car Safety Seats for

Growing Children. Guidelines for Counseling Parents (RE9618). ElkGrove Village, IL, American Academy of Pediatrics, 199634. Protect Your Kids in the Car. Washington, DC, National High-

way Traffic Safety Administration, Emergency Nurses Association,American College of Emergency Physicians, 199735. Gotschall C, Better A, Blaus D, et al: Injuries to children

restrained in 2- and 3-point belts. 42nd Annual Proceedings of theAssociation for the Advancement of Automotive Medicine, Charlottes-ville, VA, Association for the Advancement of Automotive Medicine,1998, pp 29-4336. Winston FK, Durbin DR: BUCKLE UP! is not enough: Enhanc-

ing protection of the restrained child. JAMA 281:2070-2072, 199937. Winston FK, Durbin DR, Bhatia E, et al: Patterns of inappro-

priate restraint for children in crashes. 43rd Annual Proceedings of theAssociation for the Advancement of Automotive Medicine, Barcelona,Spain, Association for the Advancement of Automotive Medicine,1999, pp 59-6938. Manary MJ, Jaffe DM: Cervical spine injuries in children.

Pediatr Ann 25:423-428, 1996


31

ORIGINAL CONTRIBUTION

The Canadian C-Spine Rule for Radiographyin Alert and Stable Trauma PatientsIan G. Stiell, MD, MSc, FRCPCGeorge A. Wells, PhDKatherine L. Vandemheen, BScNCatherine M. Clement, RNHoward Lesiuk, MDValerie J. De Maio, MD, MScAndreas Laupacis, MD, MScMichael Schull, MD, MScR. Douglas McKnight, MDRichard Verbeek, MDRobert Brison, MD, MPHDaniel Cass, MDJonathan Dreyer, MDMary A. Eisenhauer, MDGary H. Greenberg, MDIain MacPhail, MD, MHScLaurie Morrison, MD, MScMark Reardon, MDJames Worthington, MBBS

MORE THAN 1 MILLION PA-tients with blunt traumaand potential cervicalspine (C-spine) injury are

treated each year in US emergency de-partments (EDs).1,2 Among those pa-tients presenting with intact neurologi-cal status (arriving either walking or byambulance), the incidence of acute frac-ture or spinal injury is less than 1%.3-5

Due to concerns about potentially dis-abling spinal injuries, most cliniciansmake liberal use of C-spine radiogra-phy.6-9 Nevertheless, such practice is in-efficient—more than 98% of C-spine ra-diographs are negative for fracture.10-16

Furthermore, there is considerable prac-tice variation among well-trained emer-gency physicians, with radiography ratesranging as much as 6-fold.17 Cervical

spine radiography is an example of a“little ticket” item, a low-cost proce-dure that significantly adds to health carecosts due to its high volumes of use.18,19

Author Affiliations: Division of Emergency Medi-cine (Drs Stiell, Greenberg, Reardon, and Worthing-ton), Department of Medicine (Drs Stiell, Wells, andLaupacis), Department of Epidemiology and Commu-nity Medicine (Drs Stiell and Wells), Division of Neu-rosurgery (Dr Lesiuk), and Clinical Epidemiology Unit(Drs Stiell and De Maio, and Mss Vandemheen andClement), University of Ottawa, Ottawa, Ontario; De-partment of Emergency Medicine, Queen’s Univer-sity, Kingston, Ontario (Dr Brison); Division of

Emergency Medicine, University of Toronto, Toronto,Ontario (Drs Schull, Verbeek, Cass, and Morrison); Di-vision of Emergency Medicine, University of West-ern Ontario, London (Drs Dreyer and Eisenhauer); Di-vision of Emergency Medicine, University of BritishColumbia, Vancouver (Drs McKnight and MacPhail).Corresponding Author: Ian G. Stiell, MD, MSc, FRCPC,Clinical Epidemiology Unit, F6, Ottawa Health Re-search Institute, 1053 Carling Ave, Ottawa, Ontario,Canada K1Y 4E9 (e-mail: [email protected]).

Context High levels of variation and inefficiency exist in current clinical practice re-garding use of cervical spine (C-spine) radiography in alert and stable trauma patients.

Objective To derive a clinical decision rule that is highly sensitive for detecting acuteC-spine injury and will allow emergency department (ED) physicians to be more se-lective in use of radiography in alert and stable trauma patients.

Design Prospective cohort study conducted from October 1996 to April 1999, in whichphysicians evaluated patients for 20 standardized clinical findings prior to radiography.In some cases, a second physician performed independent interobserver assessments.

Setting Ten EDs in large Canadian community and university hospitals.

Patients Convenience sample of 8924 adults (mean age, 37 years) who presentedto the ED with blunt trauma to the head/neck, stable vital signs, and a Glasgow ComaScale score of 15.

Main Outcome Measure Clinically important C-spine injury, evaluated by plainradiography, computed tomography, and a structured follow-up telephone inter-view. The clinical decision rule was derived using the ! coefficient, logistic regressionanalysis, and "2 recursive partitioning techniques.

Results Among the study sample, 151 (1.7%) had important C-spine injury. The re-sultant model and final Canadian C-Spine Rule comprises 3 main questions: (1) is thereany high-risk factor present that mandates radiography (ie, age #65 years, danger-ous mechanism, or paresthesias in extremities)? (2) is there any low-risk factor pres-ent that allows safe assessment of range of motion (ie, simple rear-end motor vehiclecollision, sitting position in ED, ambulatory at any time since injury, delayed onset ofneck pain, or absence of midline C-spine tenderness)? and (3) is the patient able toactively rotate neck 45° to the left and right? By cross-validation, this rule had 100%sensitivity (95% confidence interval [CI], 98%-100%) and 42.5% specificity (95%CI, 40%-44%) for identifying 151 clinically important C-spine injuries. The potentialradiography ordering rate would be 58.2%.

Conclusion We have derived the Canadian C-Spine Rule, a highly sensitive deci-sion rule for use of C-spine radiography in alert and stable trauma patients. If pro-spectively validated in other cohorts, this rule has the potential to significantly reducepractice variation and inefficiency in ED use of C-spine radiography.JAMA. 2001;286:1841-1848 www.jama.com

See also p 1893 and Patient Page.

©2001 American Medical Association. All rights reserved. (Reprinted) JAMA, October 17, 2001—Vol 286, No. 15 1841

Downloaded From: http://jama.jamanetwork.com/ by a SCELC - Loma Linda University User on 08/12/201532

There are no widely accepted guide-lines that have been shown to be bothsafe and efficient in guiding the use ofC-spine radiography. Recently, clini-cal decision rules have been devel-oped to guide physicians in making di-agnostic or therapeutic decisions—for example, the use of radiography forpatients with ankle or knee inju-ries.20-23 A clinical decision rule may bedefined as a decision-making tool thatis derived from original research andthat incorporates 3 or more variablesfrom the history, physical examina-tion, or simple tests.24,25 The NationalEmergency X-Radiography Utiliza-tion Study (NEXUS) low-risk criteriafor C-spine radiography were recentlyevaluated in a large study of EDs thatfound the criteria to be 99.6% sensi-tive for clinically important injuries.26

However, the specificity was only12.9%, leading to concerns that use ofthe NEXUS criteria would actually in-crease the use of radiography in someUS jurisdictions and in most coun-tries outside of the United States.

We believe that the current ineffi-ciency and variability of clinical prac-tice can be remedied with the develop-ment of an accurate, reliable, andclinically sensible decision rule. Hence,the objective of this study was to de-rive a clinical decision rule that wouldbe highly sensitive for detecting acuteC-spine injury among patients sustain-ing blunt trauma who are alert andstable but at risk for neck injury. Thiswill ultimately allow physicians to bemore selective in their use of radiogra-phy without jeopardizing patient care.

METHODSStudy Setting and PopulationThis prospective cohort study was con-ducted in 10 large Canadian commu-nity and university hospitals and in-cluded consecutive adult patientspresenting to the ED after sustainingacute blunt trauma to the head or neck.We did not include the many patientspresenting with trivial injuries, such assimple lacerations to the face.The treat-ing physician’s decision of whether to or-der radiography had no bearing on the

enrollment of patients into the study. Pa-tients were eligible for enrollment if theywere at some risk for C-spine injury ei-ther because they had neck pain fromany mechanism of injury, or becausethey had no neck pain but had all of thefollowing: some visible injury above theclavicles, had not been ambulatory, andhad sustained a dangerous mechanismof injury. In addition, patients had to bealert, which was defined as a GlasgowComa Scale (GCS) score of 15 (scalerange, 3-15), and stable, defined as nor-mal vital signs (systolic blood pressure!90 mm Hg and respiratory rate be-tween 10 and 24/min).

Patients were excluded if they: (1)were younger than 16 years; (2) had mi-nor injuries, such as simple lacera-tions, and did not fulfill the first 2 in-clusion criteria above; (3) had a GCSscore lower than 15; (4) had grossly ab-normal vital signs; (5) were injuredmore than 48 hours previously; (6) hadpenetrating trauma; (7) presented withacute paralysis; (8) had known verte-bral disease (ankylosing spondylitis,rheumatoid arthritis, spinal stenosis, orprevious cervical surgery), as deter-mined by the examining physician; (9)had returned for reassessment of thesame injury; or (10) were pregnant. Eli-gible patients transferred from otherhospitals with suspected C-spine in-jury were enrolled at the study sites withthe proviso that physicians complete thedata form prior to reviewing radio-graphic films. Many of these patientsproved not to have C-spine injury. Theresearch ethics committees of the studyhospitals approved the protocol with-out the need for informed consent. Pa-tients followed up had an opportunityto give verbal consent during the tele-phone interview conducted by a studynurse.

Standardized Patient AssessmentAll patient assessments were made bystaff physicians certified in emergencymedicine or by supervised residents inemergency medicine training pro-grams. The physician assessors weretrained with a 1-hour session to evalu-ate patients for 20 standardized clini-

cal findings from the history, generalexamination, and assessment of neu-rological status. These potential pre-dictor variables were selected by a teamof investigators at a planning consen-sus conference based on a review of theexisting literature and on results of apilot study. Findings were recorded ona data collection sheet prior to radiog-raphy. A subset of patients, where fea-sible, were independently assessed bya second emergency physician to judgeinterobserver agreement. An addi-tional 5 demographic variables were ob-tained from hospital records by studynurses.

Outcome Measuresand AssessmentThe primary outcome measure wasclinically important cervical spine in-jury, defined as any fracture, disloca-tion, or ligamentous instability dem-onstrated by diagnostic imaging.Clinically unimportant cervical spine in-juries generally do not require stabiliz-ing treatment or specialized follow-upand the definition for this has been stan-dardized based on the results of a for-mal survey of 129 neurosurgeons, spi-nal surgeons, and emergency physiciansat 8 tertiary care hospitals.27 All C-spine injuries were considered clini-cally important unless the patient wasneurologically intact and had 1 of 4 in-juries: (1) isolated avulsion fracture ofan osteophyte (2) isolated fracture ofa transverse process not involving afacet joint (3) isolated fracture of a spi-nous process not involving the laminaor (4) simple compression fracture in-volving less than 25% of the vertebralbody height.

After the clinical examination, pa-tients underwent plain radiography ofthe C-spine according to the judg-ment of the treating physician, not ac-cording to any preset guidelines. Ra-diographs were interpreted by qualifiedstaff radiologists who were blinded tothe contents of the data collection sheet.The reliability of the radiography in-terpretations was assessed by having allabnormal radiographs and 1% (ran-domly selected) of normal radio-

CERVICAL SPINE RADIOGRAPHY

1842 JAMA, October 17, 2001—Vol 286, No. 15 (Reprinted) ©2001 American Medical Association. All rights reserved.

Downloaded From: http://jama.jamanetwork.com/ by a SCELC - Loma Linda University User on 08/12/201533

A Prospective Multicenter Study of Cervical Spine Injury in Children

Peter Viccellio, MD*; Harold Simon, MD‡; Barry D. Pressman, MD§; Manish N. Shah, MD!;William R. Mower, MD, PhD¶; and Jerome R. Hoffman, MA, MD¶, for the NEXUS Group

ABSTRACT. Objective. Pediatric victims of blunttrauma have developmental and anatomic characteristicsthat can make it difficult to assess their risk of cervicalspine injury (CSI). Previous reports, all retrospective innature, have not identified any cases of CSI in eitherchildren or adults in the absence of neck pain, neurologicsymptoms, distracting injury, or altered mental status.The objective of this study was to examine the incidenceand spectrum of spine injury in patients who are youngerthan 18 years and to evaluate the efficacy of the NationalEmergency X-Radiography Utilization Study (NEXUS)decision instrument for obtaining cervical spine radiog-raphy in pediatric trauma victims.

Methods. We performed a prospective, multicenterstudy to evaluate pediatric blunt trauma victims. Allpatients who presented to participating emergency de-partments underwent clinical evaluation before radio-graphic imaging. The presence or absence of the follow-ing criteria was noted: midline cervical tenderness,altered level of alertness, evidence of intoxication, neu-rologic abnormality, and presence of painful distractinginjury. Presence or absence of each individual criterionwas documented for each patient before radiographicimaging, unless the patient was judged to be too unstableto complete the clinical evaluation before radiographs.The decision to radiograph a patient was entirely at thephysician’s discretion and not driven by the NEXUSquestionnaire. The presence or absence of CSI was basedon the final interpretation of all radiographic studies.Data on all patients who were younger than 18 yearswere sequestered from the main database for separateanalysis.

Results. There were 3065 patients (9.0% of all NEXUSpatients) who were younger than 18 years in this cohort,30 of whom (0.98%) sustained a CSI. Included in thestudy were 88 children who were younger than 2, 817who were between 2 and 8, and 2160 who were 8 to 17.Fractures of the lower cervical vertebrae (C5–C7) ac-counted for 45.9% of pediatric CSIs. No case of spinal

cord injury without radiographic abnormality was re-ported in any child in this study, although 22 cases werereported in adults. Only 4 of the 30 injured children wereyounger than 9 years, and none was younger than 2 years.Tenderness and distracting injury were the 2 most com-mon abnormalities noted in patients with and withoutCSI. The decision rule correctly identified all pediatricCSI victims (sensitivity: 100.0%; 95% confidence interval:87.8%–100.0%) and correctly designated 603 patients aslow risk for CSI (negative predictive value: 100.0%; 95%confidence interval: 99.4%–100.0%).

Conclusions. The lower cervical spine is the mostcommon site of CSI in children, and fractures are themost common type of injury. CSI is rare among patientsaged 8 years or younger. The NEXUS decision instru-ment performed well in children, and its use could re-duce pediatric cervical spine imaging by nearly 20%.However, the small number of infants and toddlers inthe study suggests caution in applying the NEXUS crite-ria to this particular age group. Pediatrics 2001;108(2).URL: http://www.pediatrics.org/cgi/content/full/108/2/e20; cervical spine, radiography, injury, pediatric.

ABBREVIATIONS. CSI, cervical spine injury; NEXUS, NationalEmergency X-Radiography Utilization Study; CT, computed to-mography; MRI, magnetic resonance imaging; SCIWORA, spinalcord injury without radiographic abnormality.

Unrecognized cervical spine injury (CSI) canproduce catastrophic neurologic disability.Fear of failing to diagnose such an injury has

led to the use of radiographic spine imaging in vir-tually all multiple blunt trauma victims. In the pedi-atric population, this issue is compounded further bythe absence of any prospective studies evaluating theselection of candidates for imaging studies afterblunt trauma.

Younger children present an additional challenge,as they are developmentally unable to communicatecrucial symptoms. Furthermore, the physical exami-nation can be limited by lack of cooperation in ananxious, crying child. Anatomic differences betweenthe pediatric and adult cervical spine are prominentuntil approximately 8 years of age and persist to alesser degree until approximately 12 years of age.1–6

As a result, details of the presentation of CSI inadults are not necessarily applicable to children. Forall of these reasons, despite the rarity of CSI in chil-dren, physicians feel compelled to use liberal radio-graphic imaging to avoid missing any case of signif-icant injury.

Previous reports, all retrospective in nature, havenot identified any cases of CSI in either children or

From the *Department of Emergency Medicine, SUNY Stony Brook Univer-sity Hospital, Stony Brook, New York; ‡Department of Pediatrics, Divisionof Emergency Medicine, Emory University School of Medicine, Children’sHealthcare of Atlanta, Atlanta, Georgia; §Department of Radiology, Cedars-Sinai Medical Center, Los Angeles, California; !Department of EmergencyMedicine, University of Rochester, School of Medicine and Dentistry, Roch-ester, New York; and ¶UCLA Emergency Medicine Center, UCLA School ofMedicine, Los Angeles, California.This article was presented in part at the Society for Academic EmergencyMedicine; May 24, 2000; San Francisco, CA; and at the Pediatric AcademicSocieties and American Academy of Pediatrics Joint Meeting; May 16, 2000;Boston, MA.Received for publication Nov 20, 2000; accepted Mar 26, 2001.Reprint requests to (P.V.) Department of Emergency Medicine, SUNY StonyBrook University Hospital, L4-515, Stony Brook, NY 11794. E-mail:[email protected] (ISSN 0031 4005). Copyright © 2001 by the American Acad-emy of Pediatrics.

http://www.pediatrics.org/cgi/content/full/108/2/e20 PEDIATRICS Vol. 108 No. 2 August 2001 1 of 6

34

adults in the absence of neck pain, neurologic symp-toms, distracting injury, or altered mental status.7–18

Although this suggests that such criteria could beuseful in limiting radiography among pediatric pa-tients, limitations in the quality of such data makeany such conclusions tenuous. The National Emer-gency X-Radiography Utilization Study (NEXUS) re-cently validated a decision instrument, based on 5low-risk criteria, that allows physicians to identify asubset of patients in whom radiographic evaluationcan be avoided safely.19 In this article, a substudy ofNEXUS that deals with blunt trauma victims who areyounger than 18 years, we define the characteristicsof CSI and evaluate the performance of the NEXUSdecision instrument for all pediatric patients and invarious subgroups defined by age.

METHODSNEXUS was a prospective observational study, and radio-

graphs were ordered entirely at the discretion of the examiningphysician, dictated only by each physician’s usual and customarypractice for obtaining such films. The study protocol neither man-dated nor directed any element of patient care and thus posed norisk to patients. Waivers of informed consent were granted to eachinstitution that participated in the study. The methods of theNEXUS study have been described in detail elsewhere.20 Partici-pating institutions (see the “Appendix”) included a mix of com-munity hospitals, academic medical centers, tertiary care facilities,trauma centers, and children’s hospitals. Essential elements aresummarized below.

Patients who were selected for radiographic imaging under-went a minimum 3-view examination, including cross-table lat-eral, anteroposterior, and open-mouth odontoid views. Other im-aging studies, including but not limited to oblique views, flexion-extension radiographs, or cervical computed tomography (CT),were ordered at the discretion of the treating physician. CT ormagnetic resonance imaging (MRI) of the entire cervical spinecould be substituted for standard 3-view studies when plain filmimaging was impractical or impossible. Only those patients whounderwent radiographic evaluation are included in the NEXUSstudy. No patient who received radiographs was excluded fromthe study.

Treating physicians completed data forms that captured demo-graphic information (gender, race, date of birth), the clinical sta-bility of the patient, and the presence or absence of each of the 5low-risk criteria. The low-risk criteria were defined as the absenceof each of the following: midline cervical tenderness, evidence ofintoxication, altered level of alertness, focal neurologic deficit, andpresence of a distracting painful injury. At each participatingcenter, physicians undertook a brief training program in the ap-plication of the NEXUS criteria before initiation of the study.Physicians could indicate that they were unable to measure anindividual criterion (eg, tenderness in a comatose patient), inwhich case, the patient was considered not to have met the low-risk criterion. A patient was considered to have met the decisioninstrument classification of low-risk only when all 5 of the low-risk criteria were fulfilled.

All findings were entered into a computer data terminal, whichthen issued the voucher that was required to obtain cervical spineradiographs. If an attending physician believed that even theminimal delay associated with completing the brief NEXUS ques-

tionnaire potentially might be harmful to the patient, then he orshe could obtain the study voucher by indicating that the patientwas unstable. In such cases, clinicians were encouraged to com-plete a full assessment of all criteria as soon as feasible, preferablybefore radiograph results were known. Instability of this type wasconsidered to represent a surrogate for significant injury andexclusion from low-risk classification.

Study radiologists at each site interpreted all radiographic stud-ies. The presence or absence of CSI and unstable CSI was based onthe final interpretation of all radiographic studies (including spe-cial imaging techniques such as plain tomography, CT, and MRI).When these reports were ambiguous, study investigators re-viewed both the reports and the original radiographs to determinefinal fracture classification. Neither the official radiology interpre-tation nor the coding of injuries was done with knowledge of thefindings on the NEXUS data form.

Cases in which the victim had a CSI and was not defined aslow-risk by the decision instrument were considered to be truepositive, whereas those who were diagnosed as having an injuryin the presence of all low-risk criteria were classified as falsenegative. Cases that met all low-risk criteria and had no CSI wereclassified as true negative, whereas those without injury but fail-ing to meet at least 1 of the criteria were classified as false positive.Sensitivity and negative predictive value (as well as specificityand positive predictive value) of the decision instrument for CSIthen were calculated in the standard manner.

Results were evaluated for all pediatric study subjects, as wellas for various subgroups of children, by age. Such subgroups werebased on developmental stage, as follows: 0 to 2 years (lack ofverbal ability), 2 to 8 years (immature cervical spine), and 8 to 17years (older children with fully developed spinal anatomy).

RESULTSA total of 34 069 patients, of all ages, were enrolled

in NEXUS (see Table 1), including 3065 children(9.0%) who were younger than 18 years. The agedistribution for pediatric cases is shown in Fig 1. Agegroupings are presented on the basis of lack of verbalability (0–2), immature cervical spine (2–8), the olderpediatric population (9–17), and the adult popula-tion (18!). Thirty children were found to have a CSI,representing 0.98% of all pediatric patients and 3.7%of all injuries reported in NEXUS. No pediatric pa-tient who was defined as low risk sustained a CSI inthis study.

Compared with the adult population, a larger per-centage of the pediatric population (603 [19.7%]) wasconsidered low-risk by the decision instrument (Ta-ble 1), and none of these children had a CSI. Table 2shows that the percentage of (stable) patients whounderwent radiographic evaluation, despite beingdefined as low-risk by the decision instrument, de-creased with every increase in age group.

Applying the NEXUS criteria to the pediatric pop-ulation, sensitivity, specificity, negative predictivevalue, and positive predictive value are noted inTable 3 and compared with the results from theoverall NEXUS population. The NEXUS decision in-strument identified all pediatric patients with CSI,

TABLE 1. NEXUS Summary

Factor Pediatric Nonpediatric Total

Number of cases 3065 31 004 34 069Number of patients with CSI (%) 30 (0.98%) 788 (2.54%) 818 (2.4%)Injury rate 0.98% 2.54% 2.4%Number of low-risk patients* 603 (19.7%) 3706 (12.0%) 4309 (12.6%)Number of low-risk patients

with injury*0 8 8

* By NEXUS decision instrument.

2 of 6 THE NEXUS PEDIATRIC CERVICAL SPINE INJURY STUDY

35

so sensitivity for CSI was 100%. Nevertheless, be-cause of the small number of children with CSI, the95% confidence interval is relatively wide (87.8%–100.0%).

Table 4 compares the NEXUS findings in the 30injured and 3035 noninjured patients. Tendernessand distracting injury were the 2 most common ab-normalities noted in patients both with and withoutCSI. Evidence of intoxication was noted infrequentlyamong children, occurring in none of the injuredpatients and in only 110 of the others.

Of 30 patients with CSI, 21 were male, and theirages ranged between 2 and 17 years, with two-thirdsbeing teenagers. No CSI was present in an infantwho was younger than 2 years. CSIs among thesechildren were distributed throughout the cervicalspine, with the majority occurring in the lower cer-vical spine (Table 5).

Table 6 provides details of all 30 pediatric patientswho sustained CSI. Most of these patients (24) wereclassified as clinically stable, and only a small num-ber of the individual low-risk criteria (16 of 150[11%]) were unable to be evaluated. There was nocase of spinal cord injury without radiographic ab-normality (SCIWORA) in this series. Among the 30children with CSI, 5 (17%) had radiographic proof ofspinal cord injury. Neurologic deficit also was re-corded on the NEXUS data form in 3 other childrenwith CSI and could not be determined in 3 more.

None of the injured children was low risk by theNEXUS instrument, and there was !1 non–low-risk

finding in 13 of the 30, with an average of 1.8 positiveamong this group. (Of children without CSI, theaverage number of positive criteria was 1.4.) Nopatient was identified as non–low risk by the deci-sion instrument solely because clinicians were un-able to assess 1 or more of the criteria. Furthermore,in no case was intoxication the only non–low-riskfinding in any of these 30 children.

DISCUSSIONNo CSI was identified in the pediatric population

without at least 1 positive NEXUS risk factor. No CSIwould have been missed had the NEXUS criteriabeen applied to this population, but 20% fewer ra-diographs would have been performed. This is con-sistent with the findings of the NEXUS study,19

which validated a clinical decision instrument de-signed to identify patients who are at extremely lowrisk of CSI.20

Our study does not prove definitively that theNEXUS low-risk criteria can be applied to childrenwith complete safety. Although NEXUS was largeenough to define the overall sensitivity of the instru-ment with substantial precision, there were few chil-dren with CSI. Thus, although the decision rule cor-rectly identified all 30 children with CSI, the lowerconfidence interval for sensitivity in this group isonly 87.8%. There were only 4 injured children whowere younger than 9 years with CSI; thus, we areeven less confident about application of the NEXUSdecision instrument to that group.

Given that it would require a prospective study ofapproximately 80 000 children to be able to defineconfidence intervals for sensitivity of 0.5% (andvastly more to do so for younger children), a defin-itive answer is unlikely to be forthcoming. Neverthe-less, we believe that it is reasonable, particularly inthe adolescent age group, to endorse cautiously theuse of the NEXUS decision instrument in children,for several reasons.

First, there is not a single case in the medicalliterature of a child with an occult CSI who wouldhave been classified as low risk by the NEXUS crite-ria. Previous studies that attempted to define criteriafor obtaining cervical spine radiographs in childrenare based on retrospective chart reviews.21–24 How-ever, even in such retrospective studies, essentiallyall pediatric patients with CSI were reported to haveneck pain or tenderness, abnormal neurologic find-ings, or altered mental state. There have been noreports of CSI among children who are old enough tocommunicate and who have normal mental status,no neck pain, no neurologic abnormality, and nodistracting injury.

Second, although there are anatomic differencesbetween adults and small children, which lead todifferent types and locations of injuries,2–8 there areno clear reasons to believe that symptoms of injuryand thus the low-risk criteria would be different inthese groups, with the exception of children who aretoo small to express themselves or to localize com-plaints. Although CSI is exceedingly rare in veryyoung children (Finally, there was a substantialnumber of children in NEXUS (603) who were iden-

Fig 1. Age distribution for pediatric patients.

TABLE 2. Percentage of Study Participants, by Age Group,Classified by NEXUS Decision Instrument as Low-Risk*

Age Low Risk

n %

0–8 216/811 26.69–17 365/1977 18.518" 3187/24786 12.9Total 3768/27574* 13.7

* For stable patients whose age is known; omits patients whoseprecise age is not known.

http://www.pediatrics.org/cgi/content/full/108/2/e20 3 of 6

36

tified as low risk for CSI, none of whom proved tohave fracture. Thus, not only was the negative pre-dictive value of the decision instrument among chil-dren 100%, but also the confidence interval places thenegative predictive value at no worse than 99.4%.For all of these reasons, therefore, we believe thatcervical spine imaging rarely should be undertakenin children who clearly are negative by the low-riskcriteria. Again, given the small number of infantsand toddlers in this study, these conclusions may notbe applicable to that group.

None of the 30 children with CSI were identifiedby the presence of intoxication alone, whereas eachof the other 4 criteria was necessary for the instru-ment to have attained 100% sensitivity. This proba-bly reflects the relative rarity of intoxication amongchildren and the small number of CSI identified.Because the overall NEXUS study19 demonstratesclearly that intoxication can obscure other findings inCSI, it certainly should remain part of the decisioninstrument in children, although it will be importantonly occasionally.

CSI was present in only 0.98% of the children in

whom radiography was performed, which is lessthan half the rate seen in adults in NEXUS. This mayreflect more liberal ordering on the part of clinicians,a lower risk of injury among small children (becauseof anatomic differences or, more likely, less exposureto dangerous mechanisms of injury), or some otherundefined factor. More liberal ordering also is sug-gested by the fact that the percentage of patients whomet the low-risk criteria and for whom radiographswere ordered decreased with each increasing agegroup (Table 2). There are substantial anatomic dif-ferences in younger children (up to age 8), whereasfrom ages 8 to 12 there is a transitional period, afterwhich the cervical spine is almost fully developedand for all practical purposes is comparable to that ofan adult.2–5 This may help to explain differences insusceptibility to injury or the basis for a differentordering policy among many clinicians.

CSI did not occur among infants who wereyounger 2 years, only 88 of whom were enrolled inthe study. This is strong testimony to the rarity of CSIin this age group. It also is possible that cases ininfants, when they do occur, present differently. Be-cause injuries in this age group have been reported tooccur predominantly to the higher cervical cord,7–10

they frequently are lethal at the scene and thus neveridentified. The smaller number of cases in this agegroup is consistent with the lower incidence of ex-posure to a mechanism of injury that causes CSIcompared with other age groups.

Almost all of the pediatric CSI in NEXUS occurredin older children, and the pattern of injury in thatgroup was not dissimilar from that seen in adults.The rare cases of CSI in younger children occurredprimarily in the upper cervical spine. This confirmsobservations of most but not all earlier studies, inwhich the distribution of CSI in children who areolder than 12 years is similar to that of adults,whereas the preponderance of injuries in youngerchildren is in the C1 and C2 regions.9,10,15–17

SCIWORA has been described prominently amongchildren,7,25,26 but the only such cases in NEXUSoccurred among adults. Because for the purpose of

TABLE 3. Operator Characteristics of the NEXUS Decision Instrument, by Age

Characteristic PediatricParticipants

Overall

Sensitivity 100% (87.8%–100.0%) 99.02% (98.0%–99.6%)Negative predictive value 100% (99.2%–100%) 99.81% (99.6%–100.0%)Specificity 19.9% (18.5%–21.3%) 12.93%Positive predictive value 1.2% (0.8%–1.8%) 2.72%

TABLE 4. Prevalence of Individual Low-Risk Criteria in Children With and Without CSI

Injured (n ! 30) Uninjured (n ! 3035)

Criterion Positive Negative N/A* Positive Negative N/A

Tenderness 21 4 5 1179 1333 523Distracting injury 11 17 2 878 1915 242Altered LOC 6 21 3 520 2326 189Neurologic findings 8 19 3 176 2611 248Intoxication 0 27 3 110 2730 195Classified as stable 24 6 0 2764 271 0

LOC indicates level of consciousness; N/A, unable to be assessed.

TABLE 5. Injuries for Pediatric and Nonpediatric Patients

Injury Pediatric Nonpediatric

Spine levelOccipital condyle 1 19C1 5 90C2 (nonodontoid) 2 192Odontoid 2 90C3 0 50C4 5 79C5 9 170C6 9 233C7 10 218Cord injuries (documented) 5 64

Interspace injuries (by level)Atlanto-occipital 2 3C1-C2 0 23C2-C3 1 20C3-C4 4 19C4-C5 1 37C5-C6 5 53C6-C7 2 52C7-T1 0 9


37

this study we could confirm the presence of a cordinjury only when it was diagnosed radiographically(typically by MRI) at the participating institution, itis possible that the 5 children who were reported tohave cord injury represent an underestimate of thetrue number of children with cord injury. There wereno children, however, in whom CSI was identified,either by imaging or clinically, who had negativeplain cervical spine radiographs (and thus met thestandard definition of SCIWORA). Given that ourmethodology included mandatory data entry for all

participants, as well as institutional follow-up of riskmanagement and neurosurgical logs, it is extremelyunlikely that cases of SCIWORA occurred but werenot captured. Because all children who underwentradiography were included, it is virtually unimagin-able that a child with clinical findings of cord injurywould not have undergone radiographic evaluation.Because a number of tertiary care facilities partici-pated in the NEXUS Group, referral bias is unlikelyto explain the absence of pediatric SCIWORA. Fi-nally, our methodology did in fact identify 22 cases

TABLE 6. Individual Children With CSI

Age Gender Stable Intoxicated Tender Injury AlteredLOC

NeurologicFindings

Fracture Type

2 F Yes No N/A No Yes Yes C2 type III odontoid fracture3 M Yes No No Yes No No Occipital condyle fracture6 M No No Yes Yes N/A N/A Craniocervical dissociation8 M Yes No Yes No No No C1 anterior and posterior arch

fracture, C2 type II odontoidfracture

9 M No N/A N/A Yes Yes Yes C4 flexion tear drop fracture11 M No No N/A Yes Yes Yes Craniocervical dissociation11 F No No Yes Yes No No C7 burst fracture11 M Yes No Yes No No No C5 anterior-inferior body fracture11 M Yes No Yes No Yes No C1 lateral mass fracture12 F Yes No Yes Yes No No C2 spinous process fracture13 M Yes No Yes No No No C6 spinous process fracture14 M Yes No Yes No No No C7 wedge compression fracture14 F Yes No No No No Yes C4–C5 subluxation, C5–C6

subluxation, C5 body andposterior element fractures,C4–C6 cord contusion*

16 F No No Yes Yes No No C7 compression fracture16 M Yes No Yes No Yes Yes C6 burst fracture and bilateral

laminar fractures, C7 body andbilateral laminar fractures

16 F Yes No N/A No Yes No C6–C7 subluxation/dislocation,C7 fracture†

16 M Yes No Yes No No No C5 burst fracture and bilaterallaminar fractures; C5–C6subluxation

16 M Yes N/A N/A Yes N/A N/A C5 body fracture; C5–C6subluxation

16 M Yes N/A Yes N/A N/A N/A C5 and C6 trabecular fractures,C3–C7 interspinous ligamentinjury

16 M Yes No No No No Yes C6 articular facet fracture; C6compression fracture; C5–C6bilateral interfacetaldislocation; C5–C6 cordcontusion*

16 M Yes No Yes No No No C4 compression fracture; C3–C4subluxation; C3–C4 cordcontusion*

16 F Yes No Yes No No Yes C4 burst fracture; C4–C5subluxation; C4–C5 cordcontusion*

16 M Yes No Yes No No No C1 posterior arch fracture17 M Yes No Yes Yes No No C7 spinous process fracture17 F Yes No Yes Yes No No C7 body fracture17 M Yes No No No No Yes C6–C7 facet and capsular injury17 M Yes No Yes N/A No No C5 flexion teardrop, C5 laminar

fracture, C6 body fracture, C5–C6 interfacetal dislocation, C5–C6 cord contusion*

17 M Yes No Yes No No No C3–C4 subluxation17 M No No Yes Yes No No C1 burst fracture, C6 and C7

spinous process fractures17 F Yes No Yes No No No Axis body fracture

CSI indicates cervical spine injury; LOC, level of consciousness.* Cord injury documented by magnetic resonance imaging.† Missed by computed tomography.

http://www.pediatrics.org/cgi/content/full/108/2/e20 5 of 6

38

of SCIWORA in adults. Thus, despite the previoussporadic reports, our data, collected systematically ina very large number of consecutive patients, make itclear that SCIWORA is at most very rare amongchildren. This most likely reflects the rarity withwhich any spinal cord injury is seen in younger agegroups. This rarity is supported by a review of allcases in the National Pediatric Trauma Registry overa 5-year period (24 740 total cases), during which anaverage of 81 cases of CSI per year were reported inchildren aged 0 to 20, with fewer than 8 per year inchildren who were younger than 2.27

This prospective observational study, the largest(and only prospective) study ever done regardingpediatric CSI, provides a great deal of informationabout this entity and strongly suggests that childrenwho meet all of the NEXUS low-risk criteria gener-ally do not need to undergo cervical spine imaging.Given the small number of children with CSI in thisseries, we urge caution in applying the decision in-strument to individual patients, particularly in theyoungest age groups, in which the number of studyparticipants was relatively small. Nevertheless, ourdata suggest that CSI is at most extremely rareamong children who are defined as low risk by thedecision instrument, concordant with both the ab-sence of any report of occult fracture in the pediatricliterature and the overall results among all 34 069NEXUS participants. We believe that this stronglysupports the safety and utility of the NEXUS criteriain children, application of which could reduce cervi-cal spine imaging in children by approximately 20%,limit the time in which children are forced to remainimmobilized, and decrease their exposure to radia-tion. Finally, utilization of the NEXUS decision in-strument should lead to a substantial reduction inhealth care costs through avoidance of unnecessaryradiographs.

APPENDIX: NEXUS STUDY PARTICIPANTSThe following centers and investigators collaborated in this

study. Principal investigator: W. Mower. Co-investigator: J. Hoff-man. Steering Committee: J. Hoffman, W. Mower, K. Todd, A.Wolfson, and M. Zucker. Site investigators: Antelope Valley Med-ical Center (Los Angeles): M. Brown and R. Sisson; BellevueHospital (New York): W. Goldberg and R. Siegmann; Cedars-SinaiMedical Center (Los Angeles): J. Geiderman and B. Pressman;Crawford Long Hospital (Atlanta): S. Pitts and W. Davis; Chil-dren’s Health care of Atlanta (Atlanta): H. Simon and T. Ball;Emory University Medical Center (Atlanta): D. Lowery and S.Tigges; Grady Hospital (Atlanta): C. Finney and S. Tigges; Hen-nepin County Medical Center (Minneapolis): B. Mahoney and J.Hollerman; Jacobi Medical Center (Bronx): M. Touger, P. Gennis,and N. Nathanson; Maricopa Medical Center (Phoenix): C. Pollackand M. Connell; Mercy Hospital of Pittsburgh (Pittsburgh): M.Turturro and B. Carlin; Midway Hospital (Los Angeles): D. Kal-manson and G. Berman; Ohio State University Medical Center(Columbus): D. Martin and C. Mueller; Southern Regional Hospi-tal (Decatur): W. Watkins and E. Hadley; State University of NewYork at Stony Brook (Stony Brook): P. Viccellio and S. Fuchs;University of California, Davis, Medical Center (Sacramento): E.Panacek and J. Holmes; University of California, Los Angeles,Center for the Health Sciences (Los Angeles): J. Hoffman and M.Zucker; University of California, San Francisco, Fresno UniversityMedical Center (Fresno): G. Hendey and R. Lesperance; Univer-sity of Maryland Medical Center (Baltimore): B. Browne and S.

Mirvis; University of Pittsburgh Medical Center (Pittsburgh): A.Wolfson and J. Towers; University of Texas Health Science Cen-ter/Hermann Hospital (Houston): N. Adame, Jr., and J. Harris, Jr.

ACKNOWLEDGMENTThis work was funded by Grant R01 HS08239 from the Agency

for Healthcare Research and Quality.

REFERENCES1. Swischuk L. Anterior displacement of C2 in children: physiologic or

pathologic? Radiology. 1977;122:759–7632. Schwischuk L. The cervical spine in childhood. Curr Probl Diagn Radiol.

1984;13:1–263. Cattell HS, Filtzer DL. Pseudosubluxation and other normal variations

of the cervical spine in children. J Bone Joint Surg. 1965;47A:1295–13094. Harris JH, Mirvis SE. The normal cervical spine. In: The Radiology of

Acute Cervical Spine Trauma. 3rd ed. Baltimore, MD: Williams & Wilkins;1996:1–76

5. Fesmire FM, Luten RC. The pediatric cervical spine: developmentalanatomy and clinical aspects. J Emerg Med. 1989;7:133–142

6. Kriss VM, Kriss TC. Imaging of the cervical spine in infants. PediatrEmerg Care. 1996;13:44–49

7. Dickman C, Rekate H, Volker S, Zabramski J. Pediatric spinal trauma:vertebrae column and spinal cord injuries in children. Pedriatr Neurosci.1989;15:237–256

8. Pang D, Pollack IF. Spinal cord injury without radiographic abnormal-ity in children: the SCIWORA syndrome. J Trauma. 1989;29:654–664

9. Henrys P, Lyne E, Lifton C, Salciccioli G. Clinical review of cervicalspine injuries in children. Clin Orthop. 1977;129:172–176

10. Ruge JR, Sinson GP, McLone DG, Cerullo LJ. Pediatric spinal injury: thevery young. J Neurosurg. 1988;68:25–30

11. Anderson JM, Schutt A. Spinal injury in children. A review of 156 casesseen from 1950 through 1978. Mayo Clin Proc. 1980;55:499–504

12. McPhee IB. Spinal fractures and dislocations in children and adoles-cents. Spine. 1981;6:533–537

13. Tator CH, Edmonds VE. Acute spinal cord injury: analysis of epidemi-ologic factors. Can J Surg. 1979;22:575–578

14. Hu R, Mustard CA, Burns C. Epidemiology of incident spinal fracturein a complete population. Spine. 1996;21:492–578

15. Hill SA, Miller CA, Kosnik EJ, Hunt WE. Pediatric neck injuries. Aclinical study. J Neurosurg. 1984;60:700–706

16. Dietrich AM, Ginn-Pease ME, Bartkowski HM, King DR. Pediatriccervical spine fractures: predominantly subtle presentation. J PediatrSurg. 1991;26:995–999 (discussion 999–1000)

17. Hadley MN, Zabramski JM, Browner CM, Rekate H, Sonntag VKH.Pediatric spinal trauma. Review of 122 cases of spinal cord and verte-bral column injuries. J Neurosurg. 1988;68:18–24

18. Givens TG, Polley KA, Smith GF, Hardin WD. Pediatric cervical spineinjury: a three-year experience. J Trauma. 1996;41:310–314

19. Hoffman JR, Mower WR, Wolfson AB, Todd KH, Zucker MI, for theNational Emergency X-Radiography Utilization Study Group. Validityof a set of clinical criteria to rule out injury to the cervical spine inpatients with blunt trauma. N Engl J Med. 2000;343:94–99

20. Hoffman JR, Wolfson AB, Todd K, Mower WR. Selective cervical spineradiography in blunt trauma: methodology of the National EmergencyX-Radiography Utilization Study. Ann Emerg Med. 1998;32:461–469

21. Jaffe DM, Binns H, Radkowski MA, Barthel MJ, Engelhard HHD. De-veloping a clinical algorithm for early management of cervical spineinjury in child trauma victims. Ann Emerg Med. 1987;16:270–276

22. Baker C, Kadish H, Schunk JE. Evaluation of pediatric cervical spineinjuries. Am J Emerg Med. 1999;17:230–234

23. Rachesky I, Boyce T, Duncan B, Djelland J, Sibley B. Clinical predictionof cervical spine injuries in children. Am J Dis Child. 1987;141:199–201

24. Laham JL, Cotcamp DH, Gibbons PA, Kahana MD, Crone KR. Isolatedhead injuries versus multiple trauma in pediatric patients: do the sameindications for cervical spine evaluation apply? Pediatr Neurosurg. 1994;21:221–226

25. Leventhal HR. Birth injuries of the spinal cord. J Pediatr. 1960;56:447–453

26. Burke DC. Traumatic spinal paralysis in children. Paraplegia. 1974;11:268–276

27. Koboska ER, Keller MS, Rallo MC, Weber TR. Characteristics of pedi-atric cervical spine injuries. J Pediatr Surg. 2001;36:100–105


39

700 | VOLUME 62 | NUMBER 3 | MARCH 2008 www.neurosurgery-online.com

REVIEW

DETECTION OF PEDIATRIC CERVICAL SPINE INJURY

OBJECTIVE: In evaluating the pediatric cervical spine for injury, the use of adult pro-tocols without sufficient sensitivity to pediatric injury patterns may lead to excessiveradiation doses. Data on injury location and means of detection can inform pediatric-specific guideline development.METHODS: We retrospectively identified pediatric patients with codes from theInternational Classification of Diseases, 9th Revision, for cervical spine injury treatedbetween 1980 and 2000. Collected data included physical findings, radiographic meansof detection, and location of injury. Sensitivity of plain x-rays and diagnostic yield fromadditional radiographic studies were calculated.RESULTS: Of 239 patients, 190 had true injuries and adequate medical records; ofthese, 187 had adequate radiology records. Patients without radiographic abnormalitywere excluded. In 34 children younger than 8 years, National Emergency X-RadiographyUtilization Study criteria missed two injuries (sensitivity, 94%), with 76% of injuriesoccurring from occiput–C2. In 158 children older than 8 years, National EmergencyX-Radiography Utilization Study criteria identified all injured patients (sensitivity, 100%),with 25% of injuries occurring from occiput–C2. For children younger than 8 years,plain-film sensitivity was 75% and combination plain-film/occiput–C3 computed tomo-graphic scan had a sensitivity of 94%, whereas combination plain-film and flexion-extension views had 81% sensitivity. In patients older than 8 years, the sensitivitieswere 93%, 97%, and 94%, respectively.CONCLUSION: Younger children tend to have more rostral (occiput–C2) injuries com-pared with older children. The National Emergency X-Radiography Utilization Studyprotocol may have lower sensitivity in young children than in adults. Limited computedtomography from occiput–C3 may increase diagnostic yield appreciably in young chil-dren compared with flexion-extension views. Further prospective studies, especiallyof young children, are needed to develop reliable pediatric protocols.

KEY WORDS: Cervical spine injury, Computed tomography, flexion-extension x-rays, NEXUS, Pediatric,Sensitivity

Neurosurgery 62:700–708, 2008 DOI: 10.1227/01.NEU.0000297088.03121.D4 www.neurosurgery-online.com

Hugh J.L. Garton, M.D., M.H.Sc.Department of Neurosurgery,University of Michigan Health System,Ann Arbor, Michigan

Matthew R. Hammer, B.S.Department of Neurosurgery,University of Michigan Health System,Ann Arbor, Michigan

Reprint requests:Hugh J.L. Garton, M.D., M.H.Sc.,Department of Neurosurgery,University of Michigan Health System,1500 East Medical Center Drive,Room 3552 TC,Ann Arbor, MI 48109-0338.Email: [email protected]

Received, August 6, 2006.

Accepted, December 11, 2007.

More than 11 million children are evaluated for trau-matic injuries each year in North America (24). Manyof these injuries will raise no specter of concern for

cervical spine injury, yet concern for such an injury exists fora substantial number of them. Only a small minority of bothadult and pediatric patients is found to have true cervicalspine injury, but early recognition and treatment are critical toprevent subsequent spinal cord injury (15, 23, 28, 39).Controversy exists over how to balance radiation exposureand the cost of extensive imaging procedures with the risk ofmissing an injury in the small minority of patients. Clinicalevaluation guidelines to determine whether imaging is neces-sary to clear the cervical spine have been established foradults, but their application in pediatric patients has been con-

troversial (20, 30, 37, 38). In a large study using the NationalEmergency X-Radiography Utilization Study (NEXUS) criteriafor cervical spine evaluation, only 30 (1%) of over 3000 chil-dren were determined to have true injuries, and none demon-strated spinal cord injury without radiographic abnormalities(SCIWORA) (17, 39). In addition, fewer than 850 patients inthis population were younger than 9 years of age, and onlyfour of these young patients were injured. This left very wideconfidence intervals around the relatively high reported sensi-tivity and negative predictive value of the NEXUS instrumentwhen applied to children (39).

Anatomic and physiological differences between the pediatricand adult spine further complicate injury assessment, and radi-ographic evaluation is a much greater challenge (14, 36).

40

A P R I L 1 9 9 9 3 3 : 4 A N N A L S O F E M E R G E N C Y M E D I C I N E 4 3 7

M E T H O D O L O G Y

Methodologic Standards for the Development of

Clinical Decision Rules in Emergency Medicine

Ian G Stiell, MD, MSc, FRCPCGeorge A Wells, MSc, PhD

The purpose of this review is to present a guide to help readerscritically appraise the methodologic quality of an article orarticles describing a clinical decision rule. This guide will alsobe useful to clinical researchers who wish to answer 1 or morequestions detailed in this article. We consider the 6 majorstages in the development and testing of a new clinical decisionrule and discuss a number of standards within each stage. Weuse examples from emergency medicine and, in particular,examples from our own research on clinical decisions rules forradiography in trauma.

[Stiell IG, Wells GA: Methodologic standards for thedevelopment of clinical decision rules in emergency medicine.Ann Emerg Med April 1999;33:437-447.]

I N T R O D U C T I O N

Reports of clinical decision rules are becoming increasinglycommon throughout the medical literature and particularlywithin emergency medicine journals. Clinical decision rules(prediction rules) are designed to help physicians withdiagnostic and therapeutic decisions at the bedside. Wedefine a clinical decision rule as a decisionmaking tool thatis derived from original research (as opposed to a consen-sus-based clinical practice guideline) and incorporates 3 ormore variables from the history, physical examination, orsimple tests.1 These tools help clinicians cope with theuncertainty of medical decisionmaking and help cliniciansimprove their efficiency, an important issue as health caresystems demand more cost-effective medical practice. Arecently published example of a decision rule that helpsemergency physicians cope with uncertainty is a guidelineabout which patients with community-acquired pneu-monia require hospitalization.2 A typical example of adecision rule to improve efficiency are the Ottawa AnkleRules for the use of radiography in acute ankle injuries.3-8

Methodologic standards for the development of clinicaldecisionrules havebeen described, originally by Wasson et al9

From the Division of EmergencyMedicine, Departments of Medicineand Epidemiology and CommunityMedicine, and Ottawa Hospital Loeb Health Research Institute,University of Ottawa, Ottawa,Ontario, Canada.

Dr Stiell is a career scientist of theMedical Research Council of Canada.

Received for publication January 19, 1998. Revisions received July 6 and July 29, 1998.Accepted for publication September 11, 1998.

Address for reprints: Ian G Stiell,MD, MSc, FRCPC, ClinicalEpidemiology Unit, Ottawa Hospital Loeb Health ResearchInstitute, 1053 Carling Avenue,Ottawa, Ontario, Canada, K1Y 4E9.

Copyright © 1999 by the AmericanCollege of Emergency Physicians.

0196-0644/99/$8.00 + 047/1/94993

41

M E T H O D O L O G I C S T A N D A R D SStiell & Wells

4 3 8 A N N A L S O F E M E R G E N C Y M E D I C I N E 3 3 : 4 A P R I L 1 9 9 9

to answer 1 or more of the questions listed below. We willconsider the 6 major stages in the development and testingof a new clinical decision rule and will discuss a number ofstandards within each stage (Figure 1). We will use exam-ples from emergency medicine and, in particular, examplesfrom our own research on clinical decision rules for radio-graphy in trauma.

1 . I S T H E R E A N E E D F O R T H E D E C I S I O N R U L E ?

Clinicians should ask themselves whether there reallyappears to be a need for a particular decision rule, or whetherthe rule appears only to represent the analysis of a conve-nient set of data. Is there a demonstrated inefficiency orvariation in current medical practice, and does there appearto be the potential for improved efficiency through guide-lines or a decision rule?

Prevalence of the clinical conditionDoes the proposed decision rule deal with a clinical

problem commonly seen in emergency departments? Arule for an uncommon clinical entity is unlikely to be easilyadopted by physicians and is unlikely to contribute sig-nificantly to the overall efficiency of emergency medicinepractice. For example, because ankle injuries are one of themost common problems seen in US EDs, one might expecta decision rule for radiography to have an effect on themanagement of hundreds of thousands of patients annu-ally.12,13 We have estimated that $500 million is spentannually on ankle radiographs in North America and thateven a modest reduction in the use of ankle radiographycould lead to large health care savings. On the other hand,because injuries to the mandible are not common, a deci-sion rule for this problem is unlikely to have a significanteffect.

Current use of the diagnostic testIf the decision rule proposes to guide the ordering of a

diagnostic test, are there data clearly demonstrating thatcurrent use of the test is inefficient? For example, we havepreviously shown that 87.2% of ankle and foot radiographsand 92.4% of knee radiographs ordered in EDs are negativefor fracture.14,15 Although we have no exact data, we suspectthat a majority of hip radiographs ordered are, in fact,positive for fracture and that, consequently, hip radiographywould not be a productive area for a decision rule.

Variation in practiceIs there significant variability in clinical practice among

similar physicians or similar institutions? The low yield of

and Feinstein10 and more recently by our own researchgroup.1,11 We consider the following to be the 6 impor-tant stages in the development and testing of a fullymature decision rule. First, is there a need for a decisionrule? Has current practice been shown to be inefficient orhighly variable? Second, was the rule derived according torigorous methodologic standards? Third, has the rule beenproperly validated prospectively in a new patient group?Fourth, has the rule been successfully implemented intoclinical practice and been shown to change behavior?Fifth, would use of the rule be cost-effective according toa formal health economic analysis? Sixth, how will therule be disseminated to ensure widespread adoption inthe emergency medicine community?

The purpose of this review is to present a guide to helpreaders critically appraise the methodologic quality of anarticle or articles describing a clinical decision rule. Thisguide will also be useful to clinical researchers who wish

Figure 1.Checklist of standards for 6 stages in the development of aclinical decision rule.

1. Is there a need for the decision rule?Prevalence of the clinical conditionCurrent use of the diagnostic testVariation in practiceAttitudes of physiciansClinical accuracy of physicians2. Was the rule derived according to methodologic standards?Definition of outcomeDefinition of predictor variablesReliability of predictor variablesSelection of subjectsSample sizeMathematical techniquesSensibility of the decision ruleAccuracy3. Has the rule been prospectively validated and refined?Prospective validationSelection of subjectsApplication of the ruleOutcomesAccuracy of the ruleReliability of the rulePhysicians’ interpretation RefinementPotential effect 4. Has the rule been successfully implemented into clinicalpractice?Clinical trial Effect on useAccuracy of the ruleAcceptability5. Would use of the rule be cost-effective?6. How will the rule be disseminated and implemented?

42

Pediatric EM

Using cervical spine clearance guidelines in apediatric population: a survey of physician practicesand opinions

Emma C. Burns, BSc, MD*; Natalie L. Yanchar, MD, MSc, FRCSC*

ABSTRACT

Background: Unlike in adults, there are currently no standard-ized, validated guidelines to aid practitioners in clearing thepediatric cervical spine (C-spine). Many pediatric centres inCanada have locally produced, adult-modified guidelines, buttheextenttowhichtheseorotherguidelinesareusedisunknown.Objective: The purpose of this study was to determine ifCanadian physicians are using either locally produced or adultC-spine guidelines to clear the C- spines of patients , 16 yearsof age. The study also characterized the common methodsused by physicians to clear pediatric C-spine injuries in termsof clinical examination and radiologic imaging.Methods: A 20-question survey was distributed to 240Canadian pediatric emergency physicians and trauma teamleaders using the Dillman Total Design Method.Results: The response rate was 68%. The results showed that61% of physicians currently use guidelines to assist in theclearance of pediatric C-spines. Of those physicians not usingguidelines, 85% stated that they would use them if they wereavailable. The clinical criteria most often used to clear pediatricC-spines were a normal neurologic examination (97%) and theabsence of C-spine tenderness (95%), intoxication (94%), anddistracting injuries (87%).Conclusions: Guidelines are commonly used by Canadianphysicians when clearing the pediatric C-spine, yet few arevalidated in children. Those most commonly used are locallydeveloped guidelines, the Canadian C-spine guidelines, orNational Emergency X-Radiography Utilization Study(NEXUS) low-risk criteria.

RESUME

Contexte: Contrairement a ce qu’il en est pour les adultes, iln’existe actuellement aucune regle normative et validee pouraider les medecins a determiner quand enlever le colliercervical chez les enfants. Au Canada, bon nombre de centres

hospitaliers pour enfants utilisent leurs propres regles pouradultes modifiees, mais l’on ne sait pas dans quelle mesureces regles ou d’autres sont utilisees.Objectif: Le but de cette etude etait de determiner si lesmedecins au Canada utilisaient des regles locales ou d’autresregles pour adultes pour determiner quand enlever le colliercervical chez les patients de moins de 16 ans. L’etude aegalement caracterise les methodes couramment utilisees(examen clinique et imagerie radiologique) par les medecinspour faire cette determination.Methode: Un sondage a 20 questions a ete distribue a 240medecins d’urgence pediatrique et chefs d’equipe de trau-matologie au Canada selon la methode de conceptionglobale de Dillman.Resultats: Le taux de reponse a ete de 68 %. Les resultats ontmontre que 61 % des medecins utilisaient des regles pouraider a determiner quand enlever le collier cervical chez lesenfants. Parmi les medecins qui ne se servaient pas deregles, 85 % ont declare qu’ils le feraient si elles etaientdisponibles. Les criteres cliniques les plus souvent utilisespour determiner quand enlever le collier cervical chez lesenfants etaient un examen neurologique normal (97 %),l’absence de sensibilite a la colonne cervicale (95 %), aucuneevidence d’intoxication du patient (94 %) et l’absence deblessures distrayante (87 %).Conclusion: Les medecins au Canada utilisent courammentdes regles pour determiner quand enlever le collier cervicalchez les enfants, mais dans la majorite des cas, ces regles nesont pas validees pour les enfants. Les regles les pluscouramment utilisees sont celles elaborees localement, laRegle canadienne concernant la radiographie de la colonnecervicale et les criteres de faible risque de l’etude NEXUS(National Emergency X-Radiography Utilization Study).

Keywords: Canadian C-Spine Rule study, Cervical-Spine,guidelines, NEXUS, pediatric trauma, practice variation

From the *Department of Pediatrics, IWK Health Centre, Halifax, NS.

Correspondence to: Dr. Emma C. Burns, Department of Pediatric Emergency Medicine, Children’s Hospital of Eastern Ontario, 401 Smyth Road,Ottawa, ON K1H 8C1; [email protected]

This article has been peer reviewed.

CJEM 2011;13(1):1-6! Canadian Association of Emergency Physicians DOI 10.2310/8000.2011.100220

ORIGINAL RESEARCH N RECHERCHE ORIGINALE

2011;13(1) 1CJEM N JCMU

43

CLINICAL STUDIES

SPINAL CORD INJURY WITHOUT RADIOGRAPHICABNORMALITY IN CHILDREN, 2 DECADES LATER

Dachling Pang, M.D.Department of PediatricNeurosurgery, University ofCalifornia at Davis, Davis, andRegional Center forPediatric Neurosurgery,Kaiser Permanente Hospitals,Northern California Region,Oakland, California

Reprint Requests:Dachling Pang, M.D.,Department of PediatricNeurosurgery, Kaiser PermanenteMedical Center,280 West MacArthur Boulevard,Oakland, CA 94611.

Received, November 20, 2003.

Accepted, August 4, 2004.

OBJECTIVE: Much new research has emerged since1982, when the original descrip-tion of spinal cord injury without radiographic abnormality (SCIWORA) as a self-contained syndrome was reported. This article reviews new and old data on SCI-WORA, from the past 2 decades.METHODS: This article reviews what we have learned since 1982 about the uniquebiomechanical properties of the juvenile spine, the mechanisms of injuries, the pro-found influence of age on injury pattern and outcome, the magnetic resonanceimaging (MRI) features, and management algorithms of SCIWORA.RESULTS: The increasing use of MRI in SCIWORA has yielded ample evidence of damage invirtually all nonbony supporting tissues of the juvenile vertebral column, including rupture ofthe anterior and posterior longitudinal ligaments, intervertebral disc disruption, muscular andinterspinal ligament tears, tectorial membrane rupture, and shearing of the subepiphysealgrowth zone of the vertebral endplates. These findings provide the structural basis for thepostulated “occult instability” in the spine of a patient after SCIWORA. MRI also demonstratedfive classes of post-SCIWORA cord findings: complete transection, major hemorrhage, minorhemorrhage, edema only, and normal. These “neural” findings are highly predictive of out-come: patients with transection and major hemorrhage had profoundly poor outcome, but40% with minor hemorrhage improved to mild grades, whereas 75% with “edema only”attained mild grades and 25% became normal. All patients with normal cord signals madecomplete recovery.

The large pool of clinical data from our own and other centers also lends statistical power touphold most of our original assertions regarding incidence, causes of injury, pathophysiology,age-related changes in the malleability of the spine, vectors of deformation, and the extremevulnerability of young children to severe cord injury, particularly high cervical cord injury.Thoracic SCIWORA has been identified as an important subset, comprising three subtypesinvolving high-speed direct impact, distraction from lap belts, and crush injury by slow movingvehicles. Computation of the sensitivities of MRI and somatosensory evoked potentials indetecting SCIWORA shows that both tests were normal in 12 to 15% of children with definite,persistent myelopathy; all of these children were nevertheless braced for 3 months because oftheir clinical syndrome. Children with transient deficits but abnormal MRI and/or somatosen-sory evoked potentials were also braced, but the 60% with transient deficits and normal MRIand somatosensory evoked potentials were not braced. This is a change from our originalpolicy in 1982 of bracing all children with persistent or transient deficits, brought on by ournew MRI and electrophysiology data.CONCLUSION: Injury prevention, prompt recognition, use of MRI and electrophysi-ological verification, and timely bracing of SCIWORA patients remain the chiefmeasures to improve outcome.

KEY WORDS: Biomechanics, Electrophysiological data, Magnetic resonance imaging, Mechanism of injury,Outcome predictions, Pediatric trauma, Spinal cord injury

Neurosurgery 55:1325-1343, 2004 DOI: 10.1227/01.NEU.0000143030.85589.E6 www.neurosurgery-online.com

NEUROSURGERY VOLUME 55 | NUMBER 6 | DECEMBER 2004 | 1325

44

Clinical Neurology and Neurosurgery 110 (2008) 429–433

Review

SCIWORA in MRI eraKemal Yucesoy a,1, K. Zafer Yuksel b,∗

a Department of Neurosurgery, Faculty of Medicine, Dokuz Eylul University, Izmir, Turkeyb Department of Neurosurgery, Faculty of Medicine, Sutcu Imam University, Kahramanmaras, Turkey

Received 17 July 2007; received in revised form 4 February 2008; accepted 6 February 2008

Abstract

The aim of this report is to discuss the use of the term ‘Spinal Cord Injury Without Radiographic Abnormality’ (SCIWORA) in the medicalliterature ever since MRI became commonly employed in the diagnosis of spinal cord injuries. Using the PubMed database and the keywords‘SCIWORA and MRI’, we found 30 published articles in the English-language literature. Incidence, clinical and radiological data, and MRIfindings were evaluated in all articles, which included one meta-analysis, two reviews, 10 case series, and 17 case reports. The incidenceof SCIWORA among children was found to be between 3.3% and 32.0%. This wide range was directly related to patients’ age, authors’specialty, and utilization of MRI. After MRI became commonly used for spinal injuries, the term has taken on an ambiguous meaning in theliterature. In our opinion, if any pathology is detected on MRI with or without radiographic abnormality, the patients should not be classed,as SCIWORA and ‘real-SCIWORA’ should be determined as ‘Spinal Cord Injury Without Neuroimaging Abnormality’ in cases with normalMRI.© 2008 Elsevier B.V. All rights reserved.

Keywords: SCIWORA; Magnetic resonance imaging; CT; Plain radiographs

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4292. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4303. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4304. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4305. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 432

1. Introduction

Spinal cord injury is not common but it is a very serioushealth problem, especially during childhood [1]. Large sizedifferences between the head and neck, weakness of the par-avertebral muscles, and elasticity of the vertebral bone canall contribute to spinal cord injury without bone pathology

∗ Corresponding author at: Kahramanmaras Sutcuimam University,School of Medicine, Department of Neurosurgery, 46050 Kahramanmaras,Turkey. Tel.: +90 344 221 23 37x369; fax: +90 344 221 23 71.

E-mail address: [email protected] (K.Z. Yuksel).1 Tel.: +90 232 412 12 12.

in pediatric trauma cases [2]. Pang and Wilberger [3] usedthe term ‘SCIWORA’ for the first time to describe 28 clinicalcases that were radiographically normal but neurologicallyabnormal. Subsequent case series and reports adopted theirdescription and used the same terminology, but standardizedcriteria for the term were never developed.

Two decades after his first publication, Pang [4] reviewedSCIWORA in all its aspects, including biomechanical, clini-cal, and radiological features, and he underlined the particularimportance of MRI. In recent years, compressive spinal cordpathologies observed on MRI have been excluded from SCI-WORA, though intraneuronal spinal lesions still fall withinthe inclusion criteria [4–6].

0303-8467/$ – see front matter © 2008 Elsevier B.V. All rights reserved.doi:10.1016/j.clineuro.2008.02.004

45

430 K. Yucesoy, K.Z. Yuksel / Clinical Neurology and Neurosurgery 110 (2008) 429–433

We have noticed a great deal of confusion in the literatureabout which MRI-visible abnormalities in the spinal cord orextraneural tissues should be included in SCIWORA. Dif-ferent applications of the term from article to article cangive rise to confusion among clinicians. We suggest thatcases without any neuroimaging abnormality on all imag-ing examination modalities (including MRI) be accepted as‘real-SCIWORA’, and that the acronym ‘SCIWORA’ shouldbe relabelled as ‘Spinal Cord Injury Without NeuroimagingAbnormality’.

2. Methods

In this study, we reviewed all articles about SCIWORAand MRI which were published up to June 2007 in theEnglish-language research literature based on a search of thePubMed database. We found 30 articles comprising one meta-analysis [7], two reviews [4,8], 10 case series [1,5,6,9,10–15],and 17 case reports [2,16–31]. We evaluated the incidence,patients’ clinical and radiological data, and MRI findings.The patients’ MRI findings are summarized in Table 1. Thepatients who showed no abnormality in any radiological testsincluding MRI were accepted as ‘real-SCIWORA’. We pro-posed that the full name of SCIWORA should stand for“Spinal Cord Injury Without Neuroimaging Abnormality”.

3. Results

The incidence of SCIWORA among children was foundto be between 3.3% and 32.0% over all the articles that wereviewed [4,7,28,32,33]. Reported incidences were linked tothe specialty of the authors: they were higher in cases eval-uated by pediatricians, pediatric surgeons, and emergencydepartment doctors, compared to cases reported by neuro-surgeons or orthopedic surgeons [4,7,30,34]. In addition, thestudies that failed to use MRI adequately to carry out diag-nosis showed high SCIWORA rates [4].

When we reviewed patients’ neurological status at admis-sion, we found that ∼80% of patients with normal MRI scansshowed a good clinical outcome. Their admission Frankelscores are summarized in Table 1.

In all articles examined, MRI was used as a diagnosticinvestigation method for spinal trauma, but no consen-sus diagnostic criteria were followed in classifying casesas SCIWORA. Included in SCIWORA were spinal cordpathologies including contusion, edema, trans-section, andhematomyelia, as well as extraneural pathologies such astraumatic disc herniation, epidural hematoma, rupture ofligaments, capsules and muscles, or the presence of a brit-tle growth zone underlying the endplate. Several (105)cases presented no abnormality on MRI. We classifiedthese cases as ‘Spinal Cord Injury Without NeuroimagingAbnormality’.

4. Discussion

SCIWORA is common, especially in childhood. The lit-erature reports an incidence of 5–67% of SCIWORA casesamong pediatric spinal injuries [2,35]. In reviewing the lit-erature on this subject, we found that pediatricians, pediatricsurgeons, and emergency department doctors always reportedhigher incidence ratios than neurosurgeons or orthopedic sur-geons. In other words, the specialty of the doctor affectsthe diagnosis and hence the reported incidence rate of SCI-WORA. Most of the studies by authors of these specialtieslack MRI studies and in most cases, the diagnosis of SCI-WORA was made based on radiographic or CT studies. To besure, MRI may not be suitable at the beginning of treatment ofpoly-traumatic patients, and the variability in diagnostic cri-teria when MRI is not used can cause confusion about the realincidence of SCIWORA. The fact remains that it is unreliableto eliminate the possibility of spinal cord lesions without anMRI examination. We believe that MRI is the gold standardfor making a diagnosis of ‘real-SCIWORA’ and determiningthe prognosis of these patients. This imaging modality shouldbe applied in all patients suspected of presenting SCIWORA.

Based on a PubMed database search, 53 articles in theEnglish-language literature up to June 2007 have used theacronym SCIWORA. Only 30 of them use or mention theMRI technique. Plain radiography and computerized tomog-raphy are neither sensitive nor specific enough to detectintraneuronal lesions of the spinal cord. Therefore we believethat only the SCIWORA articles published in the ‘MRI era’can provide reliable information on this pathology.

Launay et al. reported that approximately 44% of thepatients in their meta-analysis did not recover and 39%showed complete recovery [7]. We found that some of the arti-cles that they reviewed lacked MRI examinations. Accordingto our findings, if a diagnosis of SCIWORA is establishedwith MRI and the MRI is normal (real-SCIWORA), then theprognosis of the patients may be better than that reported inthe literature.

These injuries associated with SCIWORA were firstdescribed in the literature in the early 1950s, but the acronymSCIWORA was not coined until 1982 by Pang and Wilbergerbased on radiographic and CT studies [2,3]. These methodsare not the most appropriate for clarifying the exact statusof the spinal cord. Therefore we think that the diagnosis ofSCIWORA was used more often than it should have beenin older articles that did not use MRI as a diagnostic tool.This condition came to be diagnosed differently once MRItechnology became available for routine clinical applications.Most cases that were determined as SCIWORA prior to theMRI era were found to have some lesions in the spinal cord,which MRI devices could detect because of their high res-olution [15]. In our review of the literature, we found thatneuronal and extraneuronal pathologies were also reportedin SCIWORA patients, but most of these organic lesions arecurrently classified as different conditions in modern neu-rosurgery and involve different treatment strategies. These

46

K.Yucesoy,K.Z.Yuksel/ClinicalNeurologyandNeurosurgery110(2008)429–433431Table

1M

agneticresonance

imaging

findingsofSC

IWO

RA

casespublished

inthe

English-language

medicalliterature

Author

Publicationtype

Contusion

Edem

a/infarctionD

isruptionH

emato

myelia

Disc

herniationE

piduralhematom

aN

ormalM

RI(Frankel

Scoresatadm

ission)N

o.ofcases

Osenbach

[9]C

aseseries

No

detaileddata

onM

RIfindings

31M

atsumura

etal.[16]C

asereport

1–

––

––

–1

PeterandR

ode[17]

Case

report1

––

––

––

1G

rabband

Pang[10]

Case

series–

–1

–3

–3

(1B

,2D

)7

Turgutetal.[1]C

aseseries

No

detaileddata

onM

RIfindings

11D

uprezetal.[18]

Case

report–

–1

––

––

1Pollina

andL

i[19]C

asereport

––

–1

––

–1

Gupta

etal.[5]C

aseseries

44

––

7–

–15

Trumble

andM

ysilnski[20]C

asereport

1–

––

––

–1

Kotharietal.[21]

Case

report–

2–

–2

––

4D

ai[11]C

aseseries

11

––

1–

–3

Yam

aguchi[22]C

asereport

––

––

–1

–1

Dare

etal.[12]C

aseseries

–2

––

––

17(1

A,1

B,15

D)

19H

endeyetal.[13]

Case

series11

4–

–12

––

27B

oschetal.[8]

Review

41

–3

–1

51(13

B,38

D)

60O

rhunetal.[23]

Case

report1

––

––

––

1Strohm

etal.[24]C

asereport

––

––

––

1(1

A)

1E

rgunand

Oder[25]

Case

report–

1–

––

––

1W

engeretal.[26]C

asereport

––

––

1–

–1

Izma

etal.[27]C

asereport

No

detaileddata

onM

RIfindings

3Pang

[4]R

eview5

96

7–

–23

(2B

,21D

)50

Tewarietal.[14]

Case

series13

23–

––

–4

(4D

)40

Launay

etal.[7]M

eta-analysis–

3–

22

6–

13K

oestnerandH

oek[2]

Case

report–

––

––

–1

(1B

)1

Liao

etal.[6]C

aseseries

31

2–

––

3(3

D)

9B

uldinietal.[28]C

asereport

2–

––

––

–2

Lee

etal.[29]C

asereport

––

––

––

1(1

A)

1Shen

etal.[15]C

aseseries

4–

––

––

1(1

C)

5K

alraetal.[30]

Case

report–

1–

––

––

1R

oblesetal.[31]

Case

report–

1–

––

––

1

Total30

5154

1013

288

105313

47

432 K. Yucesoy, K.Z. Yuksel / Clinical Neurology and Neurosurgery 110 (2008) 429–433

lesions are described in spinal textbooks in different chap-ters from the one including SCIWORA. In contrast, somearticles classify these lesions under the name of SCIWORA[10], whereas others advocate excluding cases with extraneu-ronal lesions from SCIWORA [12,31]. Even articles relatingto SCIWORA that were published after MRI became routinein the clinic also show a lack of consensus about which casesare real-SCIWORA and whether patients with spinal lesionsshould be diagnosed with SCIWORA or not. Some of thesearticles report the importance of MRI for determining SCI-WORA but do not themselves provide adequate MRI dataon their cases [1,9,27]. Even those articles using MRI, whichcome from different institutes, do not use a uniform protocol,e.g. a standard magnetic field strength, imaging sequence, orimaging planes. There is also no consensus about the timingof MRI examinations in patients suspected of suffering fromSCIWORA. This can give rise to diagnostic confusion anderrors even in MRI-confirmed ‘real-SCIWORA’ cases.

Pang reviewed the SCIWORA diagnosis in all its aspectstwo decades after his first publication, and concludedthat MRI does not invalidate his original acronym. Fourmechanisms of injury have been clearly implicated in thepathogenesis of SCIWORA: hyperextension, hyperflexion,distraction, and ischemia [4]. Because this clinical condi-tion reflects precisely defined injury patterns and is mostlyspecific to pediatric cases, it seems reasonable to retain theSCIWORA term.

Pang also noted the importance of two trends: (1) thegrowth in the number of literature articles related to SCI-WORA, and (2) the use of MRI. Pang noticed the confusionaround this acronym and suggested the use of an algo-rithm to make the diagnosis less ambiguous. He excludedfrom SCIWORA compressive lesions detected with MRIand reclassified them as non-SCIWORA spinal cord injury[4]. He did not exclude intraneuronal lesions such as hemor-rhage and edema, which in part can be different spinal cordpathologies with different treatment protocols. Therefore itmay be more reliable to include only cases of spinal cordinjury presenting normal MRI results in this special clinicalsituation. Along these lines, Strohm et al. mentioned ‘real-SCIWORA’ cases that show no lesion in the spinal cord inMRI images [24]. Bosch et al. pointed out the ambiguitybetween spinal cord injury and SCIWORA [8]. Based onour literature review, MRI was used to diagnose SCIWORAin 313 of cases, with only 105 showing an absence of anylesion in the spinal cord, such as contusion, hematoma, andtraumatic discal hernia. We think that only these cases shouldbe regarded as ‘real-SCIWORA’. This may help to avoid con-fusion around this acronym in the literature. It may be morereliable to separate the cases with normal radiographic andCT findings but abnormal MRI findings from cases of trueSCIWORA; the former should then be described and treatedbased on the pathology detected on MRI.

Routine MRI images can fail to spot some lesions, anddiffusion-weighted MRI may be necessary in order to revealall lesions of the spinal cord [14,15]. New MRI techniques

may also be of great importance for detecting less obvious‘real-SCIWORA’ cases. It is quite possible that as imag-ing techniques and instrumentation continue to improve, theincidence of false-negative MRI results will fall.

The acronym SCIWORET (Spinal Cord Injury WithoutRadiologic Evidence of Trauma), although common in thetextbooks, has yet to be used in the medical literature. It ismore common in adults and is seen in the context of cervicalstenosis, ankylosing spondylitis, spinal stenosis, and disc her-niation. Therefore we recommend that this term be replacedwith a new descriptive terminology such as “MR-confirmedSCIWORET,” which emphasizes the method used to excludeabnormalities. This may also prevent diagnostic confusion inthe literature.

5. Conclusion

The term SCIWORA has become ambiguous and a moreprecise terminology is needed. Spinal post-traumatic epiduralhematoma, acute disc herniation, contusion, edema, trans-section, and hematomyelia are well known in neurosurgeryas primary and secondary pathologies of spinal cord injury.We believe that all pathologies detected by MRI after aspinal trauma should be used as exclusion criteria of SCI-WORA, and that SCIWORA should be relabelled as “SpinalCord Injury Without Neuroimaging Abnormality.” In addi-tion, physicians should accept this outcome as having abetter prognosis than has been reported for SCIWORA inthe literature. A Consensus Conference, including specialistsfrom different branches (pediatrics, neurology, neuroradiol-ogy, neurosurgery, orthopedics, etc.), should be convened toredefine this acronym on the basis of a well-designed prospec-tive study.

References

[1] Turgut M, Akpinar G, Akalan N, Ozcan OE. Spinal injuries in thepediatric age group: a review of 82 cases of spinal cord and vertebralcolumn injuries. Eur Spine J 1996;5(3):148–52.

[2] Koestner AJ, Hoak SJ. Spinal cord injury without radiographic abnor-mality (SCIWORA) in children. J Trauma Nurs 2001;8(4):101–8.

[3] Pang D, Wilberger Jr JE. Spinal cord injury without radiographic abnor-malities in children. J Neurosurg 1982;57(1):114–29.

[4] Pang D. Spinal cord injury without radiographic abnormality in chil-dren, 2 decades later. Neurosurgery 2004;55(6):1325–42.

[5] Gupta SK, Rajeev K, Khosla VK, Sharma BS, Paramjit, Mathuriya SN,et al. Spinal cord injury without radiographic abnormality in adults.Spinal Cord 1999;37(10):726–9.

[6] Liao CC, Lui TN, Chen LR, Chuang CC, Huang YC. Spinal cordinjury without radiological abnormality in preschool-aged children:correlation of magnetic resonance imaging findings with neurologicaloutcomes. J Neurosurg 2005;103(1 Suppl.):17–23.

[7] Launay F, Leet AI, Sponseller PD. Pediatric spinal cord injury with-out radiographic abnormality: a meta-analysis. Clin Orthop Relat Res2005;(433):166–70.

[8] Bosch PP, Vogt MT, Ward WT. Pediatric spinal cord injury withoutradiographic abnormality (SCIWORA): the absence of occult instabil-ity and lack of indication for bracing. Spine 2002;27(24):2788–800.

48

K. Yucesoy, K.Z. Yuksel / Clinical Neurology and Neurosurgery 110 (2008) 429–433 433

[9] Osenbach RK, Menezes AH. Spinal cord injury without radiographicabnormality in children. Pediatr Neurosci 1989;15(4):168–74.

[10] Grabb PA, Pang D. Magnetic resonance imaging in the evaluation ofspinal cord injury without radiographic abnormality in children. Neu-rosurgery 1994;35(3):406–14.

[11] Dai L. Imaging diagnosis of cervical spine and spinal cord injuries inchildren. Chin J Traumatol 2001;4(4):222–5.

[12] Dare AO, Dias MS, Li V. Magnetic resonance imaging correlation inpediatric spinal cord injury without radiographic abnormality. J Neu-rosurg 2002;97(1 Suppl.):33–9.

[13] Hendey GW, Wolfson AB, Mower WR, Hoffman JR. Nationalemergency X-radiography utilization study group spinal cord injurywithout radiographic abnormality: results of the national emergencyX-radiography utilization study in blunt cervical trauma. J Trauma2002;53(1):1–4.

[14] Tewari MK, Gifti DS, Singh P, Khosla VK, Mathuriya SN, Gupta S,et al. Diagnosis and prognostication of adult spinal cord injury withoutradiographic abnormality using magnetic resonance imaging: analysisof 40 patients. Surg Neurol 2005;63(3):204–9.

[15] Shen H, Tang Y, Huang L, Yang R, Wu Y, Wang P, et al. Applica-tions of diffusion-weighted MRI in thoracic spinal cord injury withoutradiographic abnormality. Int Orthop 2007;31(3):375–83.

[16] Matsumura A, Meguro K, Tsurushima H, Kikuchi Y, Wada M, NakataY. Magnetic resonance imaging of spinal cord injury without radiologicabnormality. Surg Neurol 1990;33(4):281–3.

[17] Peter JC, Rode H. Traumatic subarachnoid-pleural fistula: casereport and review of the literature. J Trauma 1993;34(2):303–4.

[18] Duprez T, De Merlier Y, Clapuyt P, Clement de Clety S, Cosnard G,Gadisseux JF. Early cord degeneration in bifocal SCIWORA: a casereport. Spinal Cord Injury Without Radiographic Abnormalities. Pedi-atr Radiol 1998;28(3):186–8.

[19] Pollina J, Li V. Tandem spinal cord injuries without radiographicabnormalities in a young child. Pediatr Neurosurg 1999;30(5):263–6.

[20] Trumble J, Myslinski J. Lower thoracic SCIWORA in a 3-year-oldchild: case report. Pediatr Emerg Care 2000;16(2):91–3.

[21] Kothari P, Freeman B, Grevitt M, Kerslake R. Injury to the spinal cordwithout radiological abnormality (SCIWORA) in adults. J Bone JointSurg Br 2000;82(7):1034–7.

[22] Yamaguchi S, Hida K, Akino M, Yano S, Saito H, Iwasaki Y. A caseof pediatric thoracic SCIWORA following minor trauma. Childs NervSyst 2002;18(5):241–3.

[23] Orhun H, Saka G, Berkel T. Injury to the spinal cord without anyradiographic abnormality in a child. Acta Orthop Traumatol Turc2002;36(3):268–72.

[24] Strohm PC, Jaeger M, Kostler W, Sudkamp N. SCIWORA-syndrome. Case report and review of the literature. Unfallchirurg2003;106(1):82–4.

[25] Ergun A, Oder W. Pediatric care report of spinal cord injury with-out radiographic abnormality (SCIWORA): case report and literaturereview. Spinal Cord 2003;41(4):249–53.

[26] Wenger M, Adam PJ, Alarcon F, Markwalder TM. Traumatic cervicalinstability associated with cord oedema and temporary quadriparesis.Spinal Cord 2003;41(9):521–6.

[27] Izma MK, Zulkharnain I, Ramli B, Muhamad AR, Harwant S. Spinalcord injury without radiological abnormality (SCIWORA). Med JMalays 2003;58(1):105–10.

[28] Buldini B, Amigoni A, Faggin R, Laverda AM. Spinal cord injurywithout radiographic abnormalities. Eur J Pediar 2006;165(2):108–11.

[29] Lee CC, Lee SH, Yo CH, Lee WT, Chen SC. Complete recoveryof spinal cord injury without radiographic abnormality and traumaticbrachial plexopathy in a young infant falling from a 30-feet-high win-dow. Pediatr Neurosurg 2006;42(2):113–5.

[30] Kalra V, Gulati S, Kamate M, Garg A. SCIWORA—SpinalCord Injury Without Radiological Abnormality. Indian J Pediatr2006;73(9):829–31.

[31] Robles LA. Traumatic spinal cord infarction in a child: case report andreview of literature. Surg Neurol 2007;67(5):529–34.

[32] Cirak B, Ziegfeld S, Knight VM, Chang D, Avellino AM, Paidas CN.Spinal injuries in children. J Pediatr Surg 2004;39(4):607–12.

[33] Dickman CA, Zabramski JM, Hadley MN, Rekate HL, SonntagVK. Pediatric spinal cord injury without radiographic abnormali-ties: report of 26 cases and review of the literature. J Spinal Disord1991;4(3):296–305.

[34] Fesmire FM, Luten RC. The pediatric cervical spine: developmentalanatomy and clinical aspects. J Emerg Med 1989;7(2):133–42.

[35] Kriss VM, Kriss TC. SCIWORA (Spinal Cord Injury withoutRadiographic Abnormality) in infants and children. Clin Pediatr1996;35(3):119–24.

49

PEDIATRICS/ORIGINAL RESEARCH

Factors Associated With Cervical Spine Injury in Children AfterBlunt Trauma

Julie C. Leonard, MD, MPH, Nathan Kuppermann, MD, MPH, Cody Olsen, MS, Lynn Babcock-Cimpello, MD, MPH, Kathleen Brown, MD,Prashant Mahajan, MD, MPH, Kathleen M. Adelgais, MD, Jennifer Anders, MD, Dominic Borgialli, DO, MPH, Aaron Donoghue, MD, MSCE,John D. Hoyle, Jr, MD, Emily Kim, MPH, Jeffrey R. Leonard, MD, Kathleen A. Lillis, MD, Lise E. Nigrovic, MD, MPH, Elizabeth C. Powell,MD, MPH, Greg Rebella, MD, MS, Scott D. Reeves, MD, Alexander J. Rogers, MD, Curt Stankovic, MD, Getachew Teshome, MD, MPH,

and David M. Jaffe, MD, for the Pediatric Emergency Care Applied Research Network*From the Department of Pediatrics, Division of Emergency Medicine (J. C. Leonard, Jaffe) and the Department of Neurosurgery, Division of Pediatric Neurosurgery (J. R.Leonard), Washington University in St. Louis School of Medicine, St. Louis Children’s Hospital, St. Louis, MO; the Department of Emergency Medicine (Kuppermann,Kim) and Department of Pediatrics (Kuppermann), University of California, Davis School of Medicine, University of California, Davis Medical Center, Sacramento, CA

(Kuppermann); the Central Data Management and Coordinating Center (Olsen) and the Department of Pediatrics, Division of Emergency Medicine (Adelgais), Universityof Utah School of Medicine, Salt Lake City, UT; the Department of Pediatrics, Division of Emergency Medicine, University of Rochester, University of Rochester Medical

Center, Rochester, NY (Babcock-Cimpello); the Department of Pediatrics, Division of Emergency Department, George Washington School of Medicine, Children’sNational Medical Center, Washington, DC (Brown); the Department of Pediatrics, Division of Emergency Medicine (Mahajan, Stankovic), Division of Emergency

Medicine, Wayne State University, Children’s Hospital of Michigan, Detroit, MI; the Primary Children’s Medical Center, Salt Lake City, UT (Adelgais); the Department ofPediatrics, Division of Emergency Medicine, Johns Hopkins School of Medicine, Johns Hopkins Children’s Center, Baltimore, MD (Anders); the Department of

Emergency Medicine, University of Michigan, Hurley Medical Center, Flint, MI (Borgialli); the Department of Pediatrics, Divisions of Emergency Medicine and CriticalCare Medicine, University of Pennsylvania School of Medicine, Children’s Hospital of Philadelphia, Philadelphia, PA (Donoghue); the Department of Pediatrics, Division

of Emergency Medicine, Michigan State University College of Human Medicine, Helen DeVos Children’s Hospital, Grand Rapids, MI (Hoyle); the Department ofPediatrics, Division of Emergency Medicine, State University of New York at Buffalo, Women and Children’s Hospital of Buffalo, Buffalo, NY (Lillis); the Department of

Pediatrics, Division of Emergency Medicine, Harvard Medical School, Boston Children’s Hospital, Boston, MA (Nigrovic); the Department of Pediatrics, Division ofEmergency Medicine, Northwestern University Feinberg School of Medicine, Children’s Memorial Hospital, Chicago, IL (Powell); the Department of Pediatrics, Division

of Emergency Medicine, Children’s Hospital of Wisconsin, Medical College of Wisconsin, Milwaukee, WI (Rebella); the Department of Pediatrics, Division of EmergencyMedicine, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH (Reeves); the Department of Emergency Medicine, Department of Pediatrics, University of

Michigan, C.S. Mott Children’s Hospital, Ann Arbor, MI (Rogers); and the Department of Pediatrics, Division of Pediatric Emergency Medicine, University of MarylandSchool of Medicine, Baltimore, MD (Teshome).

Study objective: Cervical spine injuries in children are rare. However, immobilization and imaging for potentialcervical spine injury after trauma are common and are associated with adverse effects. Risk factors for cervical spineinjury have been developed to safely limit immobilization and radiography in adults, but not in children. The purposeof our study is to identify risk factors associated with cervical spine injury in children after blunt trauma.

Methods: We conducted a case-control study of children younger than 16 years, presenting after blunt trauma,and who received cervical spine radiographs at 17 hospitals in the Pediatric Emergency Care Applied ResearchNetwork (PECARN) between January 2000 and December 2004. Cases were children with cervical spine injury.We created 3 control groups of children free of cervical spine injury: (1) random controls, (2) age andmechanism of injury-matched controls, and (3) for cases receiving out-of-hospital emergency medical services(EMS), age-matched controls who also received EMS care. We abstracted data from 3 sources: PECARNhospital, referring hospital, and out-of-hospital patient records. We performed multiple logistic regressionanalyses to identify predictors of cervical spine injury and calculated the model’s sensitivity and specificity.

Results: We reviewed 540 records of children with cervical spine injury and 1,060, 1,012, and 702 random,mechanism of injury, and EMS controls, respectively. In the analysis using random controls, we identified 8 factorsassociated with cervical spine injury: altered mental status, focal neurologic findings, neck pain, torticollis, substantialtorso injury, conditions predisposing to cervical spine injury, diving, and high-risk motor vehicle crash. Having 1 ormore factors was 98% (95% confidence interval 96% to 99%) sensitive and 26% (95% confidence interval 23% to29%) specific for cervical spine injury. We identified similar risk factors in the other analyses.

Conclusion: We identified an 8-variable model for cervical spine injury in children after blunt trauma thatwarrants prospective refinement and validation. [Ann Emerg Med. 2011;58:145-155.]

Please see page 146 for the Editor’s Capsule Summary of this article.

Provide feedback on this article at the journal’s Web site, www.annemergmed.com.A podcast for this article is available at www.annemergmed.com.

0196-0644/$-see front matterCopyright © 2010 by the American College of Emergency Physicians.doi:10.1016/j.annemergmed.2010.08.038

*Participating centers and investigators are listed in the Appendix.

Volume 58, NO. 2 : August 2011 Annals of Emergency Medicine 145

50

[Ann Emerg Med. 2011;58:145-155.]

INTRODUCTION

Cervical spine injury occurs in fewer than 1% of childrenpresenting for trauma evaluation.1 Interventions aimed atprotecting the cervical spine during out-of-hospital transportand subsequent radiographic assessment of the cervical spineduring evaluation in the emergency department (ED) arecommon and known to be associated with adverse effects,including pain, pressure wounds, encumbered airwaymanagement and respiratory function, and exposure to ionizingradiation.2-10 More than 99% of children evaluated after traumado not have cervical spine injury and therefore may beunnecessarily exposed to these harms.

Risk stratification strategies that have been developed inadults allow clinicians to limit these potentially harmfulinterventions to those at non-negligible risk of cervical spineinjury. The best known of these rules, the National EmergencyX-Radiography Utilization Study (NEXUS) criteria11,12 and theCanadian C-spine Rule for alert and stable trauma patients13 aremore than 99% sensitive for cervical spine injury in adults.When applied prospectively, these strategies were shown tosignificantly reduce the use of spinal immobilization andradiographic clearance without missing significant cervical spineinjuries.14-19

Efforts to develop similar risk stratification strategies inchildren with blunt trauma have been limited by small sample

sizes, particularly among young children.1,20,21 Generalizationof adult-derived cervical spine injury decision rules to childrenmay be hazardous because children have age-dependentdifferences in cervical spine anatomy and injury patterns, as wellas different mechanisms of injury and abilities to reportsymptoms. There is a pressing need to develop cervical spineinjury risk stratification strategies for use in injured children.The purpose of our study was to identify risk factors associatedwith cervical spine injury in children after blunt trauma.

MATERIALS AND METHODSSelection of Participants

We conducted a retrospective case-control study in which weevaluated the medical records of children presenting to 17medical centers (study sites) in the Pediatric Emergency CareApplied Research Network (PECARN) between 2000 and2004.22,23 We obtained institutional review board approvalfrom all participating sites. Children were eligible if they wereevaluated at a study site with cervical spine radiography afterblunt trauma before 16 years of age.

Children who had cervical spine injury were designated as“cases” and were identified by query of the study site billingdatabase, using the International Classification of Diseases, 9thRevision (ICD-9) codes for cervical spine injury. These codesencompass children with injuries to the cervical vertebrae,ligaments, or spinal cord and children with spinal cord injurywithout radiographic association. Each study site investigatorconfirmed the presence of a cervical spine injury by screeningthe medical record. The principal investigator and a pediatricneurosurgeon also verified every cervical spine injury byreviewing abstracted radiology reports and spine consultationnotes.

We assigned children without cervical spine injury to controlgroups. Children with Current Procedural Terminology codesfor cervical spine radiography but without ICD-9 codes forcervical spine injury were identified as potential controls; studysite investigators confirmed the absence of cervical spine injuryby record review. We selected appropriate controls whopresented closest in time within 1 year of their assigned case.We created 3 different control groups: a random control group(“random controls”); a group matched to cases according to ageand mechanism of injury category (defined in Table 1)(“mechanism of injury controls”); and for cases receivingemergency medical services (EMS) out-of-hospital care, acontrol group matched on age who had also received EMS out-of-hospital care (“EMS controls”). For each control group, weselected up to 2 controls per case to enhance the power ofidentifying risk factors.

Analyses of matched control groups were used to assesspossible bias and confounding effects of age, mechanism ofinjury, and receipt of out-of-hospital care. Additionally, theEMS control group allowed for enhanced ability to identifyfactors observable in the out-of-hospital setting. Consistency inresults between the random, mechanism of injury, and EMScontrol group analyses would strengthen confidence in their

Editor’s Capsule Summary

What is already known on this topicClinical decision rules have been developed andvalidated for adult trauma patients to guide imagingdecisions for cervical spine injury. No such rulesexist for children.

What question this study addressedThe authors performed a case-control study andmultiple logistic regression using PediatricEmergency Care Applied Research Network(PECARN) data on children younger than 16 yearsto identify cervical spine injury predictors.

What this study adds to our knowledgeUsing 540 cases and 1,060 controls, the authorsdeveloped an 8-risk-factor model that, when allwere absent, had a sensitivity of 98% and aspecificity of 26%.

How this is relevant to clinical practiceA decision rule might reduce the amount of cervicalspine imaging in children.

Pediatric Cervical Spine Injury After Blunt Trauma Leonard et al

146 Annals of Emergency Medicine Volume 58, NO. 2 : August 2011

51

validity, whereas inconsistency would suggest possible biasedcontrol group selection.24

Data Collection and ProcessingWe adhered to standard methods of chart reviews in

emergency medicine.25 Before participation, all study personnelattended research training sessions that included review of studymaterials and procedures, as well as mock chart reviews usingstandardized medical records. Once trained, on-site researchassistants conducted structured chart reviews, and all dataabstraction was subsequently verified by study site investigator(physician) review of the medical record. Variables underconsideration as risk factors for cervical spine injury weredefined a priori and selected from previous literaturedemonstrating associations with cervical spine injury or selectedbecause of biological plausibility (Table 2).

Data were collected for each candidate risk factor from 3separate sources: the study site medical record, referring EDrecord (if applicable), and EMS out-of-hospital run sheet (ifapplicable). We abstracted data by following an explicit manualof operations, which specified using findings from the first visitfor the injury event and included a source hierarchy foridentification of findings within each medical record. The data

obtained from the study site medical record were used in allanalyses unless otherwise specified.

We performed both remote and on-site monitoring to ensureadherence to data abstraction procedures. To assess theinterrater reliability of the chart abstraction, a secondinvestigator abstracted select variables for 10% of the studysample. Interobserver agreement was assessed with the !statistic, with lower 95% confidence limit greater than 0.4denoting at least moderate agreement.26 Variables with less thanmoderate interobserver agreement were retained in the analysisfor exploratory purposes; however, the reliability of thesevariables should be interpreted cautiously.

Primary Data AnalysisWe described children with cervical spine injury and children

in each control group in terms of mean age and frequencies forsex, race, payer source, EMS out-of-hospital care, transfer froma referring ED, and mechanism of injury category. Wecalculated bivariable odds ratios for cervical spine injury and95% confidence intervals (CIs) for each candidate risk factor,using unconditional logistic regression when comparing caseswith random controls and conditional logistic regression when

Table 1. Description of the study sample.

CSI Cases,No. (%),N!540

Random Controls,No. (%),

N!1,060

MOI Controls,No. (%),N!1,012

EMS Controls,No. (%),N!702

Age, y*0 to !2 27 (5) 116 (11) 41 (4) 34 (5)2 to !8 140 (26) 318 (30) 264 (26) 173 (25)8 to !16 373 (69) 626 (59) 707 (70) 495 (71)

SexMale 344 (64) 634 (60) 620 (61) 414 (59)Female 196 (36) 426 (40) 391 (39) 288 (41)

Race*†‡

White 332 (61) 497 (47) 451 (45) 333 (47)Black 94 (17) 280 (26) 270 (27) 170 (24)Other 37 (7) 51 (5) 67 (7) 45 (6)Not documented 77 (14) 232 (22) 224 (22) 154 (22)

Payer*†‡

Commercial/government/workmen’s compensation 359 (66) 547 (52) 585 (58) 389 (55)Medicaid 124 (23) 304 (29) 242 (24) 175 (25)Self/uninsured 28 (5) 69 (7) 68 (7) 54 (8)Not documented 29 (5) 140 (13) 117 (12) 83 (12)

Transported from scene by EMS!

364 (67) 777 (73) 716 (71) 702 (100)Transfer from referring hospital

!†‡297 (55) 205 (19) 163 (16) 97 (14)

Mechanism of injury matching category*‡

Occupant of an automobile involved in an MVC 151 (28) 259 (24) 276 (27) 204 (29)Nonautomobile MVC (includes children hit by cars and crashes

involving motorcycles/all-terrain vehicles)73 (14) 218 (21) 129 (13) 185 (26)

Falls (includes falls from bikes and during sports; and diving) 193 (36) 386 (36) 368 (36) 198 (28)Other (includes other types of sport injuries and injuries involving

animals)123 (23) 197 (19) 239 (24) 115 (16)

CSI, Cervical spine injury; MOI, mechanism of injury; MVC, motor vehicle crash.*Cases significantly different from random controls at "".05 in t test or #2 test of homogeneity.†Cases significantly different from MOI controls at "".05 in t test or #2 test of homogeneity.‡Cases significantly different from EMS controls at "".05 in t test or #2 test of homogeneity.

Leonard et al Pediatric Cervical Spine Injury After Blunt Trauma


52

comparing cases with the mechanism of injury and EMS controlgroups.

To identify a parsimonious group of variables independentlyassociated with cervical spine injury, we constructed amultivariable unconditional logistic regression model with thecervical spine injury case group and the random control group,using forward variable selection. This procedure considered allpotential variables, adding individual variables with the largestscore !2 statistic to the model until no remaining variable had ascore !2 P!.05 when added to the model. Using the sameforward selection process, we constructed 2 conditional logisticregression models: (1) cervical spine injury cases compared withmechanism of injury controls, and (2) cervical spine injurypatients brought to the hospital by EMS with EMS controls.

For the unconditional model, we explored the influence ofstudy site on the model by introducing study site as a randomeffect.

The forward variable selection procedure for each of the 3models was repeated with 1,000 bootstrap samples to assess thestability of the selected risk factors. We considered a variable tobe validated as a predictor if it was identified as significant in

more than 50% of the bootstrap analyses.27 To determine theinfluence of missing data on the regression models, we fit finalconditional and unconditional models to multiple imputed datasets and re-estimated adjusted odds ratios.28

To evaluate how well the combination of risk factorsidentified in the unconditional regression modeldistinguished cases from controls, we calculated theproportion of cervical spine injury cases with at least 1 riskfactor (sensitivity of the model for cervical spine injury) andthe proportion of controls with no risk factors (specificity ofthe model). To be classified as having no risk factors, thepatient’s medical record had to have each of the factorsdocumented as absent. The presence of any factor placed thesubject in the at-risk category. Patients with otherwisemissing data were eliminated from this analysis. To estimatethe maximum sensitivity (and minimum specificity) of themodel, we repeated this analysis with positive findings fromthe transferring hospital ED record and EMS out-of-hospitalrun sheet to replace missing or negative study site findings.To further explore the performance of the unconditionalmodel, we repeated the sensitivity analysis for the subset of

Table 2. Variables under consideration for modeling risk of cervical spine injury in children.

Risk Factor Definition for Chart Abstraction

Altered mental status Glasgow Coma Scale score !15, AVPU scale (Alert, Voice, Pain, Unresponsive) !A, evidence ofintoxication, or mental status descriptions deemed by consensus panel to represent altered level ofconsciousness

Loss of consciousness History of loss of consciousness postinjuryNonambulatory Child "2 y reported as unable to ambulate postinjuryFocal neurologic findings Paresthesias, loss of sensation, motor weakness, or other neurologic finding deemed consistent with spine

injury by consensus panel (eg, priapism)Complaint of neck pain History states that the child (if "2 y) complained of neck painPosterior midline neck

tendernessPhysical examination notes neck tenderness as posterior, midline, or at a designated cervical level; or a

descriptor that consensus panel deemed consistent with posterior midline neck tendernessAny neck tenderness Any documented tenderness on physical examination of the neckTorticollis Torticollis, limited range of motion, or difficulty moving the neck noted in history or physical examinationSubstantial injury Observable injuries that are life threatening, warrant surgical intervention, or warrant inpatient observationExtremity Considered legs to hip and arms to clavicle (eg, long bone fractures, degloving injuries)Face Considered noncranial region of the head (eg, orbital, maxilla, or mandible fractures)Head Considered cranial region of the head (eg, skull fracture, intracranial hemorrhage)Torso Thorax including clavicles, abdomen, flanks, back including the spine and the pelvis (eg, rib fractures,

visceral or solid organ injury, pelvic fracture)Predisposing condition* Conditions known to predispose to CSI and that are observable on physical examination (Down syndrome,

Klippel-Feil syndrome, achondrodysplasia, mucopolysaccharidosis, Ehlers-Danlos syndrome, Marfansyndrome, osteogenesis imperfecta, Larsen syndrome, juvenile rheumatoid arthritis, juvenile ankylosingspondylitis, renal osteodystrophy, rickets, history of CSI or cervical spine surgery)

High-risk mechanismDiving DivingFall Fall from a height "10 ftHanging HangingHit by car Pedestrian, bicycle rider, or nonmotorized vehicle struck by a motor vehicleMVC Head-on collision, rollover, ejected from the vehicle, death in the same crash, or speed "55 miles/hOther MV Nonautomobile, MVC (eg, motorcycle)Axial load to any region of the

head*The impact was noted in history to be head first, any region of the head

Axial load to top of the head* The impact was noted in history to be head first, region noted to be top of headClothes-lining Injury the result of a rope, cable, or similar item exerting traction on the neck while the child is in motion

*Not evaluated for interrater reliability.



53

cases with injuries requiring stabilization (internal fixation,halo, or brace). In addition, we evaluated a model composedof only the risk factors common to all 3 regression models.

We performed all analyses with SAS/STAT software (version 9;SAS Institute, Inc., Cary, NC), using the LOGISTIC procedure.We performed multiple imputation of missing data with IVEware(Survey Research Center, University of Michigan).

RESULTSWe identified 2,395 children as potential cases (Figure). Of

these, 540 (23%) met inclusion criteria and were enrolled.Potential controls included 42,376 children, of whom 1,060met inclusion criteria and were enrolled as random controls;1,012, as mechanism of injury controls; and 702, as EMScontrols. There was very little overlap between the controlgroups, with only 3 control patients being used in more than 1control group. Descriptive characteristics of the cases andcontrol groups are presented in Table 1. Case patients weresignificantly older than random controls (mean age 10.4 yearsversus 8.9 years). Compared with all controls, case patients weremore likely to be white, have private insurance, and betransferred to the study site from a referring hospital.

Fifteen of the 19 candidate variables that were evaluated hadat least moderate interrater agreement. Variables with less thanmoderate agreement included substantial injuries to the head,face, and torso; and clothes-lining. These findings tended tohave low prevalence or required the abstractor to make asubjective judgment about the severity of the finding.

Bivariable analysis using random controls revealed 17variables significantly associated with cervical spine injury and 5

variables without significant associations (Table 3). Themultivariable analysis resulted in an 8-variable model thatincluded altered mental status, focal neurologic deficit,complaint of neck pain, torticollis, predisposing condition,substantial injury to the torso, high-risk motor vehicle crash,and diving. The random effect of study site was negligible,resulting in odds ratios and CIs equal to those in the presentedmodel, which ignored study site.

Bivariable analysis comparing children with cervical spineinjury with mechanism of injury controls revealed 13 variablessignificantly associated with cervical spine injury and 9 variableswithout significant associations (Table 3). The multivariableanalysis using mechanism of injury controls resulted in an 8-variable model that included altered mental status, focalneurologic deficit, complaint of neck pain, substantial injury tothe torso, diving, high-risk motor vehicle crash, axial load to anyregion of the head, and clothes-lining.

Bivariable analysis comparing children with cervical spineinjury who received EMS out-of-hospital care with EMScontrols revealed 13 variables significantly associated withcervical spine injury and 9 variables without significantassociations (Table 3). The multivariable analysis using EMScontrols resulted in an 8-variable model that included alteredmental status, nonambulatory patient, focal neurologic deficit,complaint of neck pain, torticollis, substantial torso injury,high-risk motor vehicle crash, and diving.

Bootstrapping validation of the multivariable analyses identifiedthe same set of significant predictors greater than 50% of the timein all the models except for high-risk motor vehicle crash, whichappeared in 45% of bootstrapped mechanism of injury models.

Figure. Subject identification.



54

All factors identified by the unconditional model and theconditional model using EMS controls remained significant whenmultiple imputed data sets were used. Only the odds ratio for axialload to any region of the head (odds ratio 1.2; 95% CI 1.0 to 1.4)was weakened in the matched analysis using the mechanism ofinjury control data set and multiple imputation for missing data.

The sensitivity and specificity of identifying cervical spineinjury defined by the presence of at least 1 factor in theunconditional model were 94% (95% CI 91% to 96%) and32% (95% CI 29% to 35%), respectively. The addition ofpositive findings from the transferring hospital ED record orEMS out-of-hospital run sheet improved sensitivity to 98%(95% CI 96% to 99%) and decreased specificity to 26% (95%CI 23% to 29%). There were no consistent injury patternsamong children with cervical spine injury who did not have anyof the risk factors identified in the unconditional model (Table4). All children with cervical spine injury not identified by themodel had normal neurologic outcomes (no cognitive, sensory,or motor deficits) at discharge.

The sensitivity of identifying children with cervical spineinjury who required neurosurgical stabilization (n!184),

defined by the presence of at least 1 factor in the unconditionalmodel, was also 94% (95% CI 90% to 97%). The addition ofpositive findings from the transferring hospital ED record orEMS out-of-hospital run sheet improved this sensitivity to 98%(95% CI 95% to "99%).

Six variables were common to all 3 models. These includedaltered mental status, focal neurologic deficit, complaint of neckpain, substantial injury to the torso, high-risk motor vehiclecrash, and diving. The sensitivity and specificity for identifyingcervical spine injury by the presence of at least 1 of these 6factors was 92% (95% CI 89% to 94%) and 35% (95% CI32% to 38%), respectively. The addition of positive findingsfrom the transferring hospital ED record or EMS out-of-hospital run sheet improved sensitivity to 97% (95% CI 95% to98%) and decreased specificity to 29% (95% CI 26% to 32%).

LIMITATIONSMost of the limitations of this study are inherent to retrospective

chart reviews and include the potential for ascertainment andsampling bias and missing data. The chart abstraction in our studywas rigorously conducted, however, and we used several measures

Table 3. Factors associated with cervical spine injury.

Odds Ratio (95% CI)

Random Controls MOI Controls* EMS Controls*

PredictorBivariableAnalysis

MultivariableModel

BivariableAnalysis

MultivariableModel

BivariableAnalysis

MultivariableModel

Altered mental status 2.0 (1.5–2.5) 3.0 (2.1–4.3) 2.6 (1.9–3.4) 3.6 (2.2–5.7) 2.7 (2.0–3.7) 3.4 (1.9–6.1)Loss of consciousness 1.4 (1.1–1.8)

†1.1 (0.9–1.4)

†1.4 (1.0–1.8)

‡ †

Nonambulatory 1.0 (0.8–1.3)†

1.3 (1.0–1.8)†

2.6 (1.6–4.4) 2.8 (1.2–6.6)Focal neurologic findings 8.1 (5.9–11.2) 8.3 (5.6–12.2) 5.7 (4.1–7.9) 5.5 (3.6–8.6) 8.5 (5.5–13.1) 8.8 (4.7–16.4)Complaint of neck pain 2.0 (1.6–2.5) 3.2 (2.3–4.4) 1.9 (1.5–2.5) 3.0 (2.1–4.4) 1.9 (1.4–2.5) 2.3 (1.4–3.8)Posterior midline neck

tenderness1.4 (1.1–1.8)

†1.3 (1.0–1.6)

‡ †1.2 (0.8–1.6)

†

Any neck tenderness 1.3 (1.1–1.7)†

1.1 (0.9–1.4)†

1.3 (1.0–1.7)†

Torticollis 1.8 (1.2–2.7) 1.8 (1.1–2.9) 2.1 (1.4–3.3)†

11.7 (3.4–39.7) 64.5 (6.9–602.6)Substantial injury: extremity 1.1 (0.7–1.5)

†1.3 (0.9–2.0)

†1.1 (0.7–1.6)

†

Substantial injury: face§

1.0 (0.6–1.7)†

0.7 (0.4–1.2)†

1.3 (0.7–2.3)†

Substantial injury: head§

1.6 (1.2–2.1)†

1.9 (1.4–2.6)†

2.0 (1.4–2.9)†

Substantial injury: torso§

1.9 (1.3–2.8) 1.9 (1.1–3.4) 3.7 (2.2–6.3) 4.3 (1.8–10.3) 2.8 (1.8–4.3) 2.6 (1.2–5.7)Predisposing condition 5.0 (1.6–16.0) 15.0 (2.9–78.0) 5.0 (1.6–15.9)

†1.5 (0.3–6.7)

†

High-risk mechanism: diving 73.3 (10.0–536.7) 73.0 (9.6–555.6) 16.3 (5.8–45.9) 15.4 (4.0–58.6) 32.0 (4.2–241.3) 74.3 (0.9–"999)High-risk mechanism: fall 0.5 (0.2–0.9)

†0.5 (0.3–1.1)

†0.5 (0.2–1.1)

†

High-risk mechanism: hanging 0.8 (0.0–10.4)‡ !

2.0 (0.0–78.0)‡ !

0.8 (0.0–10.6)‡ !

High-risk mechanism: hit by car 0.5 (0.4–0.7)†

0.6 (0.3–1.5)†

0.6 (0.4–0.8)†

High-risk mechanism: MVC 1.7 (1.3–2.3) 2.5 (1.8–3.6) 6.6 (2.5–17.0) 2.8 (1.0–8.3)¶

2.1 (1.5–2.8) 3.6 (2.1–6.1)High-risk mechanism: other MV 1.1 (0.6–2.0)

†1.4 (0.6–3.2)

†0.9 (0.4–1.7)

†

Axial load to any region of thehead

1.6 (1.3–2.0)†

1.5 (1.2–1.9) 1.5 (1.0–2.2)#

1.5 (1.1–2.1)‡ †

Axial load to top of the head 2.4 (1.4–4.2)†

3.2 (1.7–5.8)†

6.0 (2.2–16.5)†

Clothes-lining§

3.0 (1.2–7.5)†

2.9 (1.1–7.5) 3.0 (1.0–9.4) 4.0 (0.7–21.8)†

MV, Motor vehicle.*Conditional logistic regression was used for EMS and MOI control groups.†Not selected for inclusion in model.‡Exact estimate and CI.§! Statistic lower bound less than 0.4.!Hanging was not included in model selection because of nonprevalence in cases and less than 0.5% prevalence in controls.¶Not validated with bootstrapping.#95% CI includes 1.0 when estimated with multiple imputed data.



55

to limit these biases. These measures included uniform training ofall study personnel, explicit instructions for data abstraction foreach variable, interrater reliability measurements, and careful studymonitoring. We also used multiple control groups to assesssampling bias and multiple imputation analyses to explore theeffects of missing data.

Additionally, we identified factors by using a forward selectionprocedure that allows the entry of a new variable into the model,provided the new model is significantly improved. Because forwardselection procedures only add variables, it is possible for the finalmodel to contain variables that are significant when added but areno longer significant when considered in the presence ofsubsequently added variables. Although this did not occur forfactors in the unconditional model, CIs for the high-risk motorvehicle crash, axial load to any region of the head, and clothes-

lining odds ratios had a lower limit of 1.0 in the mechanism ofinjury model, and CIs for the diving odds ratio had a lower limit of0.9 in the EMS model.

DISCUSSIONIn this large, multicenter case-control analysis, we identified

8 factors associated with cervical spine injury in children whoexperienced blunt trauma (altered mental status, focalneurologic deficits, complaint of neck pain, torticollis,substantial injury to the torso, predisposing condition, high-riskmotor vehicle crash, and diving). These historical and physicalexamination findings are highly predictive of cervical spineinjury in children after trauma and differ somewhat frompreviously established adult screening criteria and those fromsmaller pediatric studies (Table 5).1,11-13,20,21

Table 4. Characteristics of children with CSI who did not have one of the 8 factors in the unconditional model.

Age,Years Mechanism of Injury Injury Disposition Treatment

11 children with CSImissed when alldata sourcesconsidered

5 Collision or fall from bicycle Atlantoaxial rotary subluxation Floor Rigid collar1 Fall from elevation C1 lateral mass fracture OR Brace

12 Fall from elevation C5 compression fracture Home Soft collar9 Fall from elevation Os odontoideum with ADI !5

mmHome Internal fixation*

15 Motorized transport crash (eg, ATV) C5-7 spinous process fractures Floor Rigid collar12 Sports injury C7 transverse process fracture Home Rigid collar14 Collision or fall from bicycle C2 vertebral body fracture Floor None10 Fall from elevation C3 lateral mass fracture Floor None

2 Fall down stairs SCIWORA Floor Brace10 Fall from standing/walking/running Odontoid fracture, type 2 ICU Halo12 Bicycle struck by moving vehicle C6 compression fracture ICU Rigid collar

33 children with CSImissed when onlystudy site dataconsidered

14 Collision or fall from bicycle Odontoid fracture, type 2 ICU Halo13 Motorized transport crash (eg, ATV) C6 vertebral body fracture ICU None

8 Pedestrian struck by movingvehicle

C2 lateral mass fracture ICU Rigid collar

14 Pedestrian struck by movingvehicle

C7 transverse process fracture OR Rigid collar

13 Collision or fall from bicycle SCIWORA Floor Rigid collar12 Collision or fall from bicycle C3 burst fracture with spinal

cord injuryFloor Halo

14 Fall down stairs C5 compression fracture Home Rigid collar11 Fall from elevation Os odontoideum with spinal

cord injuryICU Internal fixation

13 Sports injury C2-3 subluxation Home Rigid collar15 Collision or fall from bicycle C2 laminar fracture Floor None15 Sports injury SCIWORA Floor Rigid collar10 Blunt injury to the head/neck Hangman’s fracture Floor Halo14 Collision or fall from bicycle C5 tear drop fracture with spinal

cord injuryFloor Brace

9 Sports injury Odontoid fracture, type 3 Floor Halo1 Fall from elevation Jefferson fracture Floor Rigid collar

14 Sports injury C7 spinous process fracture Home Soft collar12 Sports injury SCIWORA Floor Rigid collar

5 Fall down stairs Odontoid fracture, type 2 Floor Halo11 Sports injury SCIWORA Floor None

6 Fall from elevation C2 spinous process fracture Floor None14 Motorized transport crash (eg, ATV) C2 spinous process fracture Floor Rigid collar12 Fall from standing/walking/running SCIWORA ICU None

SCIWORA, Spinal cord injury without radiographic association.*Discharged home with subsequent outpatient surgery.



56

The NEXUS collaborative reported a 5-variable decision rulethat was derived and validated in a predominantly adultcohort.1,11,12 Our model of cervical spine injury in childrencontains 3 of these 5 variables: altered mental status,intoxication (included in our definition of altered mentalstatus), and focal neurologic deficits. Cervical spine injuries areknown to be associated with head injuries, which is likely due tothe association with axial load as a causal biomechanical forcefor both. Additionally, individuals with acute injuries to theupper cervical cord may experience respiratory compromise,hypoxic brain injury, and subsequent altered mental status.Focal neurologic findings, although uncommon, are fairlyspecific for spinal cord injuries.

Posterior midline neck tenderness, which was important inthe NEXUS criteria, was not identified in our model. Instead,our model contains self-reported neck pain and torticollis. Weconsidered substantial injuries that were observable on physicalexamination to be chart-ascertainable proxies for the painfuldistracting injury variable described by NEXUS. Wesubcategorized substantial injuries by region of the body, and inour model, only substantial injuries to the torso were importantpredictors of cervical spine injury in children. In contrast toNEXUS, which relied solely on clinical variables, we found 2

mechanisms of injury to be important cervical spine injurypredictors in children: high-risk motor vehicle crash and diving.

The Canadian C-spine Rule is another decision rule forclinical clearance of the cervical spine in adult patients afterblunt trauma.13 Seven of the 8 factors identified in our modelare consistent with this rule. The Canadian C-spine Rule doesnot include associated injury variables such as substantial torsoinjury. Predisposing condition, a factor absent from theNEXUS criteria, is included in both our model and theCanadian C-spine Rule. These conditions, in particular Downsyndrome in children and ankylosing spondylitis in adults,although uncommon, are known to be associated with cervicalspine injury.29,30 The Canadian C-spine Rule, however,contains factors absent from our model, including falls greaterthan 3 feet or 5 stairs, crashes involving bicycles or motorizedrecreational vehicles, and inability to ambulate postinjury.Inability to ambulate, however, is a variable in our model ofcervical spine injury generated with the EMS control group.

Two small, single-center studies identified risk factors forcervical spine injury in children. One included several variablesthat were similar to those in our model: altered mental status,focal neurologic findings, complaint of neck pain, andtorticollis.20 Unlike our model, that study included general neck

Table 5. PECARN model compared with previous cervical spine injury models: a comparison of predictive variables.

Multicenter Studies Single-Center Studies

PECARN Model, NEXUS Criteria,1,11,12 Canadian C-spine Rule,13 Jaffe,20 Rachesky,21

Children With CSI in Study Sample n!540 n!30 n!0 n!59 n!25

Mental statusAltered mental status X X X* XHistory of head trauma XIntoxication X

†X

Focal neurologic deficits X XAbnormal reflexes XStrength X* XSensation XParesthesias XNeck findingsHistory of neck trauma XComplaint of neck pain X X X XTorticollis X X XGeneral neck tenderness XPosterior midline neck tenderness X XOther examination findingsPainful distracting injury XSubstantial torso injury XPredisposing condition X X*Inability to ambulate XMechanisms of injuryHigh-risk MVC X X

‡X

‡

Diving X XAxial load to the head XFall from an elevation !1 m or 5 stairs XMotorized recreation vehicle XBicycle collision X

*Considered at risk a priori and therefore excluded from derivation cohort.†Included in definition of altered mental status.‡Varies in definition when compared to PECARN definition.



57

tenderness but did not include any mechanistic factors. Anotherstudy proposed a 2-variable model (complaint of neck pain andmotor vehicle crash with associated head trauma) that was ableto identify all 25 children with cervical spine injury.21

Although 6 of the 8 risk factors for cervical spine injury weresimilar across all control groups, supporting the findings of theunconditional model, there were some different risk factorsidentified in the conditional models. Predisposing condition wasnot included in the models derived with the mechanism ofinjury and EMS control groups; however, this was one of theleast prevalent findings in our study sample. Torticollis was notincluded in the model derived with mechanism of injurycontrols, which suggests that torticollis may be related toparticular mechanisms of injury. Nonambulatory after injurywas included in the model derived with EMS age-matchedcontrols, which suggests that this factor may be important inidentifying cervical spine injury in children who receive out-of-hospital care.

The mechanism of injury-matched analyses identifiedbiomechanical forces (clothes-lining, axial load) and subsetsof motor vehicle crash (high-risk motor vehicle crash) thatwere predictive of cervical spine injury for subjects within thesame mechanism of injury-matching category. Thishighlights the importance of biomechanics and severitymarkers in defining risk factors for cervical spine injury.These risk factors, however, warrant prospective refinementbecause they were the weakest of the risk factors in themechanism of injury-matched analysis.

This study represents a large investigation of cervical spineinjury in children derived from primary source data.Although there were subtle differences between theconditional and unconditional models, the overallconsistency between the models and the bootstrappingvalidation support the stability of the unconditional model.Application of this model as a decision rule within thissample of imaged children would have detected 98% ofchildren with cervical spine injury and reduced exposure tospinal immobilization and ionizing radiation for the non–cervical spine injury children by more than 25%.

We identified 8 predictors of cervical spine injury in childrenafter blunt trauma, including altered mental status, focalneurologic deficits, complaint of neck pain, torticollis,substantial torso injury, predisposing condition, diving, andhigh-risk motor vehicle crash. These factors should be highlyconsidered in the development of a decision rule for theidentification of children at negligible risk for cervical spineinjury after blunt trauma, in whom immobilization andradiographic evaluation can be deferred.

The authors acknowledge the site principal investigators andresearch coordinators in PECARN (please see the Appendix), whosededication and hard work made this study possible, and theextraordinary work of statistician Cody Olsen, MS, from the

PECARN Central Data Management and Coordinating Center.

Supervising editors: Kelly D. Young, MD, MS; Steven M.Green, MD

Author contributions: JCL and DMJ conceived the study andobtained grant funding. JCL, NK, and DMJ designed the study.JCL, NK, LB-C, KB, PM, KA, JA, DB, AD, JDH, EK, KL, LEN, EP,GR, DMJ, SDR, AJR, CS, and GT acquired data and providedsupervision for the study. JCL and JRL verified all cervicalspine injuries. JCL, NK, CO, and DMJ conducted the dataanalysis and interpreted the data. JCL drafted the article, andall authors critically revised it. JCL takes responsibility for thepaper as a whole.

Funding and support: By Annals policy, all authors are requiredto disclose any and all commercial, financial, and otherrelationships in any way related to the subject of this articleas per ICMJE conflict of interest guidelines (seewww.icmje.org). This work was supported by a grant from theHealth Resources and Services Administration/Maternal andChild Health Bureau (HRSA/MCHB), Emergency MedicalServices of Children (EMSC) Program (H34 MC04372). ThePediatric Emergency Care Applied Research Network (PECARN)is supported by cooperative agreements U03MC00001,U03MC00003, U03MC00006, U03MC00007, andU03MC00008 from the EMSC program of the MCHB, HRSA,US Department of Health and Human Services.

Publication dates: Received for publication May 10, 2010.Revision received August 6, 2010. Accepted for publicationAugust 27, 2010. Available online October 29, 2010.

Presented at the Pediatric Academic Societies annualmeeting, May 2009, Baltimore, MD; and the Society ofAcademic Emergency Medicine annual meeting, May 2009,New Orleans, LA.

Address for correspondence: Julie C. Leonard, MD, MPH,Campus Box 8116, St. Louis Children’s Hospital, OneChildren’s Place, St. Louis, MO 63110; 314-454-2341, fax314-454-4345; E-mail [email protected].

REFERENCES1. Viccellio P, Simon H, Pressman BD, et al. A prospective

multicenter study of cervical spine injury in children. Pediatrics.2001;108:e20.

2. Chan D, Goldberg R, Tascone, A, et al. The effect of spinalimmobilization on healthy volunteers. Emerg Med Serv. 1994;23:48-51.

3. Cordell WH, Hollingsworth JC, Olinger ML, et al. Pain and tissueinterface pressures during spine board immobilization. Ann EmergMed. 1995;26:13-16.

4. Linares H, Mawson A, Suarez E, et al. Association betweenpressure sores and immobilization in the immediate post-injuryperiod. Orthopedics. 1987;10:571-573.

5. Heath KJ. The effect on laryngoscopy of different cervical spineimmobilization techniques. Anaesthesia. 1994;49:843-845.

6. Schafermeyer RW, Ribbeck BM, Gaskins J, et al. Respiratoryeffects of spinal immobilization in children. Ann Emerg Med.1991;20:1017-1019.



58

7. March JA, Ausband SC, Brown LH. Changes in physicalexamination caused by use of spinal immobilization. PrehospEmerg Care. 2002;6:421-424.

8. Broder J, Fordham LA, Warshauer DM. Increasing utilization ofcomputed tomography in the pediatric emergency department,2000-2006. Emerg Radiol. 2007;14:227-232.

9. Jimenez RR, Deguzman MA, Shiran S, et al. CT versus plainradiographs for evaluation of c-spine injury in young children: dobenefits outweigh risks? Pediatr Radiol. 2008;38:635-644.

10. Berrington de Gonzalez A, Mahesh M, Kim KP, et al. Projectedcancer risks from computed tomographic scans performed in theUnited States in 2007. Arch Intern Med. 2009;169:2071-2077.

11. Hoffman JR, Schriger DL, Mower WR, et al. Low-risk criteria forcervical-spine radiography in blunt trauma: a prospective study.Ann Emerg Med. 1992;21:1454-1460.

12. Hoffman JR, Mower WR, Wolfson AB, et al. Validity of a set ofclinical criteria to rule out injury to the cervical spine in patientswith blunt trauma. National Emergency X-Radiography UtilizationStudy Group. N Engl J Med. 2000;343:94-99.

13. Stiell IG, Wells GA, Vandemheen KL, et al. The Canadian C-spineRule for radiography in alert and stable trauma patients. JAMA.2001;286:1841-1848.

14. Kerr D, Bradshaw L, Kelly AM. Implementation of the CanadianC-Spine Rule reduces cervical spine x-ray rate for alert patientswith potential neck injury. J Emerg Med. 2005;28:127-131.

15. Stiell IG, Clement CM, Grimshaw J, et al. Implementation of theCanadian C-Spine Rule: prospective 12 centre cluster randomisedtrial. BMJ. 2009;339:b4146.

16. Muhr D, Seabrook DL, Wittwer LK. Paramedic use of a spinalinjury clearance algorithm reduces spinal immobilization in theout-of-hospital setting. Prehosp Emerg Care. 1999;3:1-6.

17. Stroh G, Braude D. Can an out-of-hospital cervical spineclearance protocol identify all patients with injuries? an argumentfor selective immobilization. Ann Emerg Med. 2001;37:609-615.

18. Domeier RM, Swor RA, Evans RW, et al. Multicenter prospectivevalidation of prehospital clinical spinal clearance criteria.J Trauma. 2002;53:744-750.

19. Armstrong BP, Simpson HK, Crouch R, et al. Prehospitalclearance of the cervical spine: does it need to be a pain in theneck? Emerg Med J. 2007;24:501-503.

20. Jaffe DM, Binns H, Radkowski MA, et al. Developing a clinicalalgorithm for early management of cervical spine injury in childtrauma victims. Ann Emerg Med. 1987;16:270-276.

21. Rachesky I, Boyce WT, Duncan B, et al. Clinical prediction ofcervical spine injuries in children. Radiographic abnormalities.Am J Dis Child. 1987;141:199-201.

22. Pediatric Emergency Care Applied Research Network. ThePediatric Emergency Care Applied Research Network (PECARN):rationale, development, and first steps. Acad Emerg Med. 2003;10:661-668.

23. Dayan P, Chamberlain J, Dean JM, et al. The Pediatric EmergencyCare Applied Research Network: progress and update. ClinPediatr Emerg Med. 2006;7:128-135.

24. Hayden GF, Kramer MS, Horwitz RI. The case-control study: apractical review for the clinician. JAMA. 1982;247:326-31.

25. Gilbert EH, Lowenstein SR, Koziol-McLain J, et al. Chart reviewsin emergency medicine research: where are the methods? AnnEmerg Med. 1996;27:305-308.

26. Landis JR, Koch GG. The measurement of observer agreement forcategorical data. Biometrics. 1977;33:159-174.

27. Chen CH, George SL. The bootstrap and identification ofprognostic factors via Cox’s proportional hazards regressionmodel. Stat Med. 1985;4:39-46.

28. Newgard CD, Haukoos JS. Advanced statistics: missing data inclinical research—part 2: multiple imputation. Acad Emerg Med.2007;14:669-678.

29. Pizzutillo PD, Herman MJ. Cervical spine issues in Downsyndrome. J Pediatr Orthop. 2005;25:253-259.

30. Chapman J, Bransford R. Geriatric spine fractures: an emerginghealthcare crisis. J Trauma. 2007;62(6 suppl):S61-62.

APPENDIXParticipating centers and investigators are listed below in alpha-

betical order.Boston Children’s HospitalBoston, MA

Michelle Betances, MPHLise E. Nigrovic, MD, MPH

State University of New York, BuffaloBuffalo, NY

Kathleen Lillis, MDHaiping Qiao, MD, MPH

Children’s Hospital of MichiganDetroit, MI

Elizabeth B. Duffy, MAPrashant Mahajan, MD, MPHCurt Stankovic, MD

Children’s Hospital of PhiladelphiaPhiladelphia, PA

Aaron Donoghue, MD, MSCEEileen Houseknecht, RNMarlena Kittick, MPH

Children’s National Medical CenterWashington, DC

Kathleen Brown, MDBobbe Thomas, MPH

Cincinnati Children’s Hospital Medical CenterCincinnati, OH

Matthew Krugh, BAScott D. Reeves, MDRegina Taylor, MA

DeVos Children’s Hospital/Spectrum HealthGrand Rapids, MI

John D. Hoyle, Jr, MDJeffery Trytko, MS

Hurley Medical CenterFlint, MI

Dominic Borgialli, DO, MPHMichael Hadden, BS

Johns Hopkins Medical CenterBaltimore, MD

Jennifer Anders, MDBrook Hanna, BSErin Van Wagener, BS

Medical College of Wisconsin and Children’s Hospital ofWisconsin

Milwaukee, WIGreg Rebella, MDDuke Wagner, DC



59

Chicago Memorial/NorthwesternChicago, IL

Elizabeth Powell, MD, MPHPrimary Children’s Medical CenterSalt Lake City, UT

Kathleen Adelgais, MDKammy Jacobsen, EMT

UC Davis Medical CenterSacramento, CA

Emily Kim, MPHNathan Kuppermann, MD, MPHShari Nichols, CCRP

University of MichiganAnn Arbor, MI

Rachel L. McDuffie, MPHAlexander J. Rogers, MD

University of Rochester Medical CenterRochester, NY

Lynn Babcock-Cilmpello, MD, MPHGeorge O’Gara, MBA

University of MarylandBaltimore, MD

Corey Bhogte, MDGetachew Teshome, MD, MPH

Washington University and St. Louis Children’s HospitalSt. Louis, MO

David M. Jaffe, MDVirginia Koors, MSPHJeffrey R. Leonard, MDJulie C. Leonard, MD, MPH

Central Data Management and Coordinating Center/Universityof Utah

Salt Lake City, UTKym Call, BAJ. Michael Dean, MD, MBARene Enriquez, BSRichard Holubkov, PhDCody Olsen, MSBen Yu, MSSally Jo Zuspan, RN, MSN

PECARN Steering CommitteeN. Kuppermann, Chair; E. Alpern, D. Borgialli, K. Brown,

J. Chamberlain, J. M. Dean, G. Foltin, M. Gerardi, M. Gore-lick, J. Hoyle, D. Jaffe, C. Johns, K. Lillis, P. Mahajan, R.Maio, S. Miller (deceased), D. Monroe, R. Ruddy, R. Stanley,M. Tunik, A. Walker. MCHB/EMSC liaisons: D. Kavana-ugh; Central Data Management and Coordinating Center(CDMCC): M. Dean, R. Holubkov, S. Knight, A. Donaldson,S. Zuspan; Feasibility and Budget Subcommittee: T. Singh,Chair; A. Drongowski, L. Fukushima, M. Shults, J. Suhajda,M. Tunik, S. Zuspan; Grants and Publications Subcommittee:M. Gorelick, Chair; E. Alpern, G. Foltin, R. Holubov, J.Joseph, S. Miller (deceased), F. Moler, O. Soldes, S. Teach;Protocol Concept Review and Development Subcommittee:D. Jaffe, Chair; A. Cooper, J. M. Dean, C. Johns, R. Kanter, R.Maio, N. C. Mann, D. Monroe, K. Shaw, D. Treloar; QualityAssurance Subcommittee: R. Stanley, Chair; D. Alexander, J,Burr, M. Gerardi, R. Holubkov, K. Lillis, R. Ruddy, M.Shults, A. Walker; Safety and Regulatory Affairs Subcommit-tee: W. Schalick, Chair; J. Brennan, J. Burr, J. M. Dean, J.Hoyle, R. Ruddy, T. Singh, D. Snowdon, J. Wright.



60

Absence of clinical findings reliably excludes unstable cervicalspine injuries in children 5 years or younger

Diane F. Hale, MD, Colleen M. Fitzpatrick, MD, John J. Doski, MD, Ronald M. Stewart, MD,and Deborah L. Mueller, MD, San Antonio, Texas

BACKGROUND: Increased accessibility and rapidity of computed tomography (CT) have led to increased use and radiation exposure to pediatrictrauma patients. The thyroid is radiosensitive and therefore at risk for developing malignancy from radiation exposure duringcervical spine CT. This analysis aimed to determine which preelementary trauma patients warrant cervical spine CT bydefining incidence and clinical characteristics of preelementary cervical spine injury.

METHODS: This was a retrospective review of pre-elementary trauma patients from 1998 to 2010 with cervical spine injury admitted toa Level I trauma center. Patients were identified from the trauma registry using DRG International Classification ofDiseasesV9th Rev. codes and reviewed for demographics, mechanism of injury, clinical presentation, injury location, injurytype, treatment, and outcome.

RESULTS: A total of 2,972 preelementary trauma patients were identified. Twenty-two (0.74%) had confirmed cervical spine injuries.Eleven (50%) were boys, and the mean (SD) age was 3 (1.7) years. The most common mechanism of injury was motor vehiclecollision (n = 16, 73%). The majority (59%) were in extremis, and 12 (55%) arrived intubated. The median Glasgow ComaScale (GCS) score was 3 (interquartile range, 3Y10); the median Injury Severity Score (ISS) was 33 (interquartile range,17Y56). Nineteen injuries (76%) were at the level of C4 level and higher. The mortality rate was 50%. All patients had clinicalfindings suggestive of or diagnostic for cervical spine injury; 18 (82%) had abnormal neurologic examination result, 2 (9%)had torticollis, and 2 (9%) had neck pain.

CONCLUSION: The incidence of cervical spine injury in preelementary patients was consistent with previous reports. Missing a cervical spineinjury in asymptomatic preelementary patients is extremely low. Reserving cervical spine CT to symptomatic preelementarypatients would decrease unnecessary radiation exposure to the thyroid. (J Trauma Acute Care Surg. 2015;78: 943Y948.Copyright * 2015 Wolters Kluwer Health, Inc. All rights reserved.)

LEVEL OF EVIDENCE: Therapeutic study, level IV.KEY WORDS: Cervical spine injury; pediatric trauma; pediatric cervical spine; radiation exposure.

During the last 20 years, the use of computed tomography(CT) has increased dramatically, from less than 10 million

per year in the early 1980s to currently greater than 62 millionper year. The largest increase in CT use has been in the cate-gories of pediatric diagnosis and adult screening.1 In a single-institution emergency department analysis during a 6-yearperiod, the pediatric population had a dramatic increase inCT use without a concomitant increase in patient acuity.2 Themost dramatic rise was in the use of cervical spine and chest

CT. In adult blunt trauma patients, increased use of cervicalspine and chest CT has evolved secondary to their superioritycompared with traditional radiographs for screening of spinefractures and blunt aortic injury. In contrast to adults, children,especially the very young (e5 years old), rarely sustain theseinjuries. The incidence of cervical spine injury in this age groupis 0.4%, while the incidence of blunt thoracic aortic injury waszero in reviews of the National Trauma Data Bank.3,4

Preelementary trauma patients (e5 years old) with cer-vical spine injury present with injury patterns different fromthose of older children and adults. They are more likely to haveinjuries to the upper cervical spine, ligamentous injuries asopposed to fractures, severe and complete spinal cord injuries,and higher mortality rates.5 Most deaths are associated withbrain injuries in young children.6 These differences are at-tributable to differences in the anatomy of the immature spineand its response to deformational forces. The acceleration anddeceleration forces at high-impact speeds cause more stress onthe upper cervical spine of young children versus older childrenand adults because of their proportionally larger head-to-bodyratio.7 The fulcrum is shifted higher to the C2Y3 level in youngchildren versus C5Y6 in older children and adults.7,8

The sharp increase in CT use raises concern for theamount of radiation exposure and its lifelong impact on thepediatric population. Children are at greater risk of developing

ORIGINAL ARTICLE

J Trauma Acute Care SurgVolume 78, Number 5 943

Submitted: November 10, 2014, Revised: January 2, 2015, Accepted: January13, 2015.

From the San Antonio Military Medical Center, Fort Sam Houston, (D.F.H.); andUniversity of Texas Health Science Center (J.J.D., R.M.S., D.J.M.), SanAntonio, Texas; and Saint Louis University School of Medicine (C.M.F.), St.Louis, Missouri.

This study was presented at the American College of Surgeons’ Committee onTrauma resident paper competition March 23, 2013, in San Diego, California,and at the American Association for the Surgery of Trauma 2012 poster pre-sentation, September 13, 2012, in Kauai, Hawaii.

The views expressed herein are those of the authors and do not reflect the officialpolicy or position of Brooke Army Medical Center, the US Army MedicalDepartment, the US Army Office of the Surgeon General, the Department of theArmy, Department of Defense, or the US Government.

Address for reprints: Deborah Mueller, MD, University of Texas HealthScience Center, 7703 Floyd Curl Drive, San Antonio, TX 78229; email:[email protected].

DOI: 10.1097/TA.0000000000000603

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.61

a malignancy from a given dose of radiation because they areinherently more radiosensitive related to ongoing growth anddevelopment and have more remaining years of life for po-tential cancer development. CT has increased radiation expo-sure compared with a simple radiograph; a neonatal abdominalCT has 2,000 times more radiation exposure than a simplechest radiograph.1 Cervical spine CT results in significant ra-diation exposure to the thyroid gland, which is known to beradiosensitive. Previously, the amount of radiation exposurewas estimated from historical data.9 In a study of prospectivelymeasured radiation exposure in 197 pediatric trauma patients,Mueller et al.10 found that thyroid doses in 71% of pediat-ric trauma patients fell within the dose range historically cor-related with an increased risk of thyroid cancer (10Y90 mGy).The average thyroid radiation exposure was approximately30 mGy; this is double the estimated exposure by previousauthors. The study also demonstrated that selective scanning ofbody areas resulted in a statistically significant decrease in alldoses when compared with whole-body scanning.

The purpose of this study was to determine the incidenceand characteristics of preelementary pediatric cervical spineinjuries at a single institution. This study aimed to determinewhich preelementary-aged trauma patients gain additionaldiagnostic utility from cervical spine CT with the goal ofavoiding unnecessary radiation in the remaining population.

PATIENTS AND METHODS

This was an institutional review boardYapproved retro-spective review of preelementary (e5 years old) trauma pa-tients from 1998 to 2010 with cervical spine injury presentingto a Level I trauma center. Patients were identified from thetrauma registry using DRG International Classification ofDiseasesV9th Rev. codes. The trauma registry, electronicmedical records, and autopsy reports were reviewed for de-mographics, mechanism of injury, clinical presentation, injurylocation, injury type, treatment, and outcome. Descriptive

statistics were calculated for sex, age, Glasgow Coma Scale(GCS) score, Injury Severity Score (ISS), presenting symp-toms, mechanism of injury, injury pattern, imaging modality,disposition, and mortality. Statistical analysis was performedusing IBM SPSS Statistics for Windows version 19.0 softwarepackage (IBM Corp., Armonk, NY).

RESULTS

Overall, 5,905 pediatric trauma patients (e16 years old)were admitted to this single institution in 12 years. A subsetof 2,972 preelementary patients (e5 years old) were identi-fied. Twenty-two patients (0.74%) had cervical spine injuryconfirmed via radiographs or autopsy (Fig. 1). Eleven patients(50%) were male, and the mean (SD) agewas 3 years (1.7). Thepreponderance of patients with an abnormal neurologic ex-amination result (n = 18) arrived with a GCS score of 3 (n = 15,83%). Three patients had GCS scores of 8, 11, and 12, ac-cording to the modified GCS for infants and children. Spe-cifically, an infant patient with a GCS score of 8 (E1V3M4) hadan absence of movement in the lower extremities and requiredintubation for respiratory distress. A 4-year-old patient with aGCS score of 11 (E4V2M5) localized to pain briskly only in hisleft upper extremity. Only one patient with an abnormal GCSscore had a normal motor score on arrival, but this patient onlyopened eyes to verbal response and was inconsolable andcrying, with a GCS score of 12 (E3V3M6).

The median ISS was 33 (interquartile range [IQR],17Y56; ISS 9 33; n = 13). The majority (59%) were in extremis,

Figure 1. Consort diagram.

TABLE 1. Demographics

Demographics (n = 22)

Age, mean (SD), y 3 (1.7)Male 11 (50%)GCS score on arrival to ED, median (IQR) 3 (3Y10)ISS, median (IQR) 33 (17Y56)Intubated on/before arrival to ED 12 (55%)Extremis on arrival 13 (59%)Overall mortality 11 (50%)

ED, emergency department; GCS, glasgow coma scale; IQR, interquartile range; ISS,injury severity index; SD, standard deviations.

Figure 2. Associated injuries.

J Trauma Acute Care SurgVolume 78, Number 5Hale et al.

944 * 2015 Wolters Kluwer Health, Inc. All rights reserved.


and 12 (55%) arrived intubated to the emergency department.The mortality rate was 50% (Table 1). Six patients expiredshortly after arrival to the emergency department; the re-maining five died shortly after admission to the intensive careunit. Of the survivors, seven were discharged home and four toacute rehabilitation facilities. Cervical spine injuries in survi-vors were treated by surgery (n = 3) and stabilization via collaror halo (n = 8). Two patients treated with cervical collars weretransferred to tertiary hospitals near their homes on hospitalday 18 and 25 and may have undergone additional treatment.Among the nine patients with follow-up visits at our facility,two had persistent neurologic deficits.

The most common mechanism of injury was motor ve-hicle collision (MVC, n = 16, 73%), followed by pedestrianstruck by vehicle (n = 4) and fall (n = 2). Most patients hadassociated injuries, suggesting high-energy impacts (Fig. 2).There was only one patient following MVC without associatedinjuries; however, this patient arrived in extremis, intubatedwith a GCS score of 3, and died shortly after arrival. Atlan-tooccipital dislocation was identified on autopsy.

All patients who did not arrive with lethal injuries hada cervical spine CT obtained (n = 17). During the 12-yearperiod of this review, CT scanners used included a 4-slice

scanner in the radiology department from 1998 to 2000, asingle-slice scanner installed in the emergency departmentfrom 2000 to 2005, followed by a 16-slice scanner from 2006 to2010. Additional imaging modalities obtained in these patientsincluded neck plain films (n = 3) and magnetic resonanceimaging (MRI, n=7). Eight patients had two imaging modal-ities, and one patient had three imaging modalities. Five did notreceive any imaging; they arrived with lethal injuries and weredeclared dead before imaging was obtained. Injury type inthese patients was determined at autopsy.

Nineteen injuries (76%) were at the level of C4 and higher(Fig. 3). There were six patients with injuries to the lowercervical spine; their median age was 2 years, with a medianGCS score of 5.5 and a median ISS of 55.5, and had a mortalityrate of 33% (n = 2). Both deaths were associated with braininjuries. One patient had both an upper and lower cervical spineinjury. There were two patients with a GCS score of 15 and anISS of 17; these two were the only patients not intubated. Bothhad neck pain and other injuries requiring admission to thehospital; both were treated with stabilizing collars (Table 2).

All patients with unstable cervical spine injury had cli-nical findings suggestive of or diagnostic for cervical spineinjury; 18 (82%) had abnormal neurologic examination result,

Figure 3. Location and type of injuries. Illustration credit: Kelly J. Rosso, MS, MD

TABLE 2. Lower Cervical Spine Characteristics

Cervical Spine Injury at C5YC7 (n = 6)

Patient Age SexMechanismof Injury GCS ISS Mortality Injuries Radiology Study Level Injury

A 0 Male MVC 8 54 Alive Spleen, pulmonary contusion,clavicle fracture

CT, MRI,plain film

C7 Cord contusion, ligament

B 1 Female MVC 3 75 Death Brain, liver/kidney contusion,pulmonary contusion

None C5Y6 Dislocation withcord transection

C 1 Male Pedestrianvs. car

3 57 Death Brain, liver, pulmonary contusion,rib fracture

CT C7YT1 Dislocation withcord epidural

D 3 Female MVC 15 17 Alive Liver laceration, rib fracture CT C7 Transverse process fractureE 4 Male Fall 15 17 Alive Adrenal hematoma CT C5YT2 Spinous process fractureF 5 Male MVC 3 75 Alive Pulmonary contusion CT, MRI C1Y2, C5Y6 Ligament distraction

spinal cord injury

J Trauma Acute Care SurgVolume 78, Number 5 Hale et al.

* 2015 Wolters Kluwer Health, Inc. All rights reserved. 945


2 (9%) had torticollis, and 2 (9%) had neck pain. In those whowere clinically evaluable (not in a coma), there were noasymptomatic patients who later were found to have unstablecervical spine injury. A subset (n = 82, 2.7%) of the 2,972patients 5 years or younger who sustained blunt trauma duringa 6-month period in 2009 were analyzed for the use of cervicalspine CT. Overall, 42% of the patients (n = 35) had evaluationwith cervical spine CT.More than half of these patients (n = 21,60%) had normal neurologic examinations, no neck tenderness,and absence of torticollis. A post hoc power analysis comparingthe known incidence reported at 0.4%3 and our observed inci-dence at 0.74%, with our sample size of 2,972, revealed that ourobserved power was 0.76.

DISCUSSION

In adult trauma patients who do not meet clinical criteriafor cervical spine clearance, CT of the cervical spine has be-come the standard imaging modality.11 Reports of increasedcervical spine CT use in pediatric patients suggest the adultpractice guideline is being used in more children despite cleardifferences in anatomy, clinical presentation, and injury pat-terns. Because of the long-term risk of radiation exposure inpediatric patients, cervical spine CT should be selectively used.The goal of this study was to document the incidence of cer-vical spine injury in preelementary trauma patients.

The incidence (0.74%) and pattern of cervical spine in-jury in the youngest trauma patients in our series was consistentwith previous reports.3,6,12Y15 Many articles have been pub-lished on pediatric spine injury including reviews of the Na-tional Trauma Data Bank and the National Pediatric TraumaRegistry. The large number of patients in these databasesmakes their use attractive when studying rare injuries such ascervical spine injury in the pediatric population; however,symptoms and physical examination findings (other than GCS)are not captured in these studies.3,6,16,17 This is a major gap inthese types of studies because clinical examination is essentialto making the clinical diagnosis of cervical spine injuries. Inaddition, these reviews do not provide details with regard tospecific fracture types or treatment required. Early adoption ofan electronic medical record at our institution allowed us toretrieve these data.

In the preelementary age group of children, our findingscall into question the need for cervical spine CT scan in theevaluation of not intubated, noncomatose children withoutclinical signs or symptoms. This is an important finding be-cause this age group of patients may be nonverbal and they arethe group most sensitive to radiation exposure.

We reviewed previous publications, which includedclinical examination data, to see whether our results are con-sistent with preelementary trauma patients from other centers.Viccellio et al.18 reported on 3,065 pediatric patients enrolledin the National Emergency X-Radiography Utilization Study(NEXUS) with more than 500 patients in the range of 5 years oryounger. No cervical spine injury was identified without atleast one positive NEXUS criterion, but this study includedonly 30 patients with cervical spine injuries in the entire groupand only 2 injuries in patients 5 years or younger. Four majorpediatric trauma centers reviewed their combined cervical

spine injuries following fall from less than 5 ft in preelementarypatients.19 Only eight children were identified with cervicalspine injuries, but every one of them had neck pain or limitedrange of neck motion. In a smaller retrospective study of theNEXUS criteria in 187 pediatric patients with cervical spineinjury and adequate records for review, Garton et al.20 reportedtwo patients with significant cervical spinal injuries withNEXUS low-risk criteria. The authors report a sensitivity of94% in this group of patients. It is not clear whether the twopatients with fractures were clinically asymptomatic with re-spect to neck range of motion. One additional case reportidentified a 3-year-old restrained passenger in a high-speedMVC with two deaths on the scene and significant seat beltmarks on physical examination but NEXUS criteria negative.21

Based on current data and these previous reports, it seemsthat there would be a very low risk of missing an unstablecervical spine injury in a preelementary-aged patient if theyare asymptomatic.

Despite our data, if routine cervical spine injuryscreening is still deemed necessary in this patient populationbecause of concerns about the ability to clinically evaluate thepediatric cervical spine, other authors have documented suf-ficient sensitivity and specificity of plain films, which delivermuch smaller amounts of ionizing radiation.22,23 Hernandez etal.22 evaluated cervical spine trauma in the pediatric populationyounger than 5 years and found that cervical spine CTwas oflow yield in identifying a positive and clinically significantfinding. They concluded that a plain film of the cervical spinewould identify the same injury pattern. Brockmeyer et al.23

evaluated the high-risk population of the obtunded child withpossible cervical spine injury and compared four diagnosticimaging modalities (plain cervical spine radiographs, flexion-extension radiograph, CT cervical spine, and MRI). Theyfound no difference in sensitivity and specificity of plain films(100% and 95%, respectively) and CT cervical spine (100%and 95%, respectively) in identifying unstable cervical spineinjuries in this high-risk population but still recommended CTto detect critical injuries in the obtunded child. The TraumaAssociation of Canada recently published a consensus guide-line that also supports choosing plain films as a first line overcervical spine CT to limit radiation exposure if radiographicevaluation is undertaken.24 The guideline recommends neckplain films with cervical spine CT only for pediatric patientswith an unreliable clinical examination finding. The UnitedKingdom also has guidelines on cervical spine evaluation in thepediatric trauma patient and advocates plain films initially, withCTas an adjunct to evaluate areas of concern.25 Kulaylat et al.26

support the use of clearing an awake and conscious patient withlow suspicion of unstable cervical spine injury by starting withplain films of the neck using a technique of cephalic stabili-zation to obtain high-quality films.

The University of Iowa implemented a protocol to clearthe pediatric cervical spine in children younger than 10 years.The protocol allowed clinical clearance of the cervical spine inthe asymptomatic alert patient without distracting injuries andobservation of normal neck range of motion. Their protocolincorporated a limited cervical spine CT to C3 if a head CTwasobtained but used plain films if a head CT was not beingobtained. After implementation of this protocol, they were able




to significantly reduce the amount of cervical spine CT scansperformed. This technique allowed for the evaluation of theportion of the cervical spine most likely injured in patients5 years or younger, thereby limiting radiation to the thyroid.27

Given this information, our institution does not routinelyobtain screening cervical spine CT in children 5 years oryounger presenting after trauma. Pediatric patients with ab-normal neurologic examination result, decreased mental status,neck pain, or torticollis are evaluated with cervical spine CT;however if the child is asymptomatic defined by a normalneurologic examination result, appropriate mental status, withabsence of neck pain or torticollis, our first step is to remove thecervical collar. We examine the patient for cervical tendernessif they are able to communicate and observe the child fornormal range of motion of the neck. In preverbal patients, wesimply observe neck range of motion with the collar removed.If the child seems to move his or her neck without discomfortand full range of motion, then we do not pursue any furtherradiologic evaluation. If the child seems reluctant to move hisor her neck, we reapply the collar and pursue radiologicworkup. Based on similar sensitivity and specificity with cer-vical spine CT and plain films, workup can be initiated withplain films.22Y26

The retrospective nature of this study is its largest lim-itation. There is a potential for missed injuries in dischargedpatients who may have received subsequent care at an outsidehospital. As our facility was the only Level 1 trauma center inthe region accepting pediatric trauma patients, the chance ofthis occurring is unlikely as the surrounding hospitals transferchildren identified with cervical spine injury to our facility forsubspecialty care. We chose the age bracket of 0 year to 5 yearsbased on developmental changes in the spine and because asubstantial portion of this subset is developmentally preverbaland because this group has the greatest risk of radiation ex-posure. Other authors have grouped children from 0 year to8 years together, but the fulcrum of the spine begins shiftingand is considered complete at age 9 years, sowe did not want toinclude older children because results may not be consistent.

CONCLUSION

The incidence of cervical spine injury in preelementarypatients was 0.76%, consistent with previous reports. There is alow likelihood of a missed unstable cervical spine injury inpediatric patients 5 years or younger with the absence ofclinical findings. Given the anatomic differences and differingdistribution of traumatic cervical spine injuries in pre-elementary children, this demographic should be consideredseparately from older children and adult trauma patients.Clinical guidelines to obtain cervical spine CT imaging forchildren 5 years or younger include neck pain, torticollis,neurologic deficit, or obtunded mental state. If a child does nothave these symptoms and there is still a concern for cervicalspine injury, then plain films of the neck are an appropriateinitial imaging modality. In the preelementary trauma popu-lation, limiting cervical spine CT to a diagnostic modality inonly symptomatic patients or patients with altered mental status(intubation or obtunded) will decrease unnecessary radiation

exposure while identifying all clinically significant cervicalspine injuries.

AUTHORSHIP

D.F.H. and D.L.M. performed the literature search. D.L.M., D.F.H.,C.M.F., and R.M.S. were responsible for the study concept and design.D.L.M. performed the data collection. D.F.H. and D.L.M. completedthe data analysis. D.F.H., D.L.M., R.M.S., C.M.F., and J.J.D. were re-sponsible for the data interpretation, drafting of the manuscript, andcritical revision.

ACKNOWLEDGMENT

We thank the research and trauma program staff of the UTHSCSA De-partment of Trauma and Emergency Surgery.

DISCLOSURE

The authors declare no conflicts of interest.

REFERENCES1. Brenner DJ, Hall EJ. Computed tomographyVan increasing source of

radiation exposure. N Engl J Med. 2007;357:2277Y2284.2. Broder J, Fordham LA,Warshauer DM. Increasing utilization of computed

tomography in the pediatric emergency department, 2000Y2006. EmergRadiol. 2007;14:227Y232.

3. Mohseni S, Talving P, Branco BC, Chan LS, Lustenberger T, Inaba K, BassM, Demetriades D. Effect of age on cervical spine injury in pediatricpopulation: a National Trauma Data Bank review. J Pediatr Surg.2011;46:1771Y1776.

4. Heckman SR, Trooskin SZ, Burd RS. Risk factors for blunt thoracic aorticinjury in children. J Pediatr Surg. 2005;40:98Y102.

5. Klimo P Jr, Ware ML, Gupta N, Brockmeyer D. Cervical spine trauma inthe pediatric patient. Neurosurg Clin N Am. 2007;18:599Y620.

6. Kokoska ER, Keller MS, RalloMC,Weber TR. Characteristics of pediatriccervical spine injuries. J Pediatr Surg. 2001;36:100Y105.

7. Khanna G, El-Khoury G. Imaging of cervical spine injuries of childhood.Skeletal Radiol. 2007;36:477Y494.

8. Kreykes NS, Letton RW Jr. Current issues in the diagnosis of pediatriccervical spine injury. Semin Pediatr Surg. 2010;19:257Y264.

9. Kim PK, Zhu X, Houseknecht E, Nickolaus D, Mahboubi S, Nance ML.Effective radiation dose form radiologic studies in pediatric traumapatients. World J Surg. 2005;29:1557Y1562.

10. Mueller DL, Hatab M, Al-Senan R, Cohn SM, Corneille MG, Dent DL,Michalek JE,Myers JE,Wolf SE, Stewart RM. Pediatric radiation exposureduring the initial evaluation for blunt trauma. J Trauma. 2011;70:724Y731.

11. Como JJ, Diaz JJ, Dunham CM, Chiu WC, Duane TM, Capella JM,Holevar MR, Khwaja KA, Mayglothling JA, Shapiro MB, et al. Practicemanagement guidelines for identification of cervical spine injuries fol-lowing trauma: update from the Eastern Association for the Surgeryof Trauma Practice Management Guidelines Committee. J Trauma.2009;67:651Y659.

12. Bilston LE, Brown J. Pediatric spinal injury type and severity are age andmechanism dependent. Spine (Phila Pa 1976). 2007;32(21):2239Y2347.

13. Platzer P, Jaindl M, Thalhammer G, Dittrich S, Kutscha-Lissberg F, VecseiV, Gaebler C. Cervical spine injuries in pediatric patients. J Trauma.2007;62:389Y396.

14. Patel JC, Tepas JJ 3rd, Mollitt DL, Pieper P. Pediatric cervical spine in-juries: defining the disease. J Pediatr Surg. 2001;36:373Y376.

15. Hofbauer M, Jaindl M, Hochtl LL, Ostermann RC, Kdolsky R, Aldrian S.Spine injuries in polytraumatized pediatric patients: characteristics andexperience from a Level I trauma center over two decades. J Trauma AcuteCare Surg. 2012;73:156Y161.

16. Polk-Williams A, Carr BG, Blinman TA, Masiakos PT, Wiebe DJ, NanceML. Cervical spine injury in young children: a National TraumaData Bankreview. J Pediatr Surg. 2008;43:1718Y1721.

J Trauma Acute Care SurgVolume 78, Number 5 Hale et al.

* 2015 Wolters Kluwer Health, Inc. All rights reserved. 947


17. Piretti-Vanmarcke R, Velmahos GC, Nance ML, Islam S, Falcone RA Jr,Wales PW, Brown RL, Gaines BA, McKenna C, Moore FO, et al. Clinicalclearance of the cervical spine in blunt trauma patients younger than3 years: a multi-center study of the American Association for the Surgeryof Trauma. J Trauma. 2009;67(3):543Y549.

18. Viccellio P, Simon H, Pressman BD, Shah MN, Mower WR, Hoffman JR;NEXUSGroup. A prospective multicenter study of cervical spine injury inchildren. Pediatrics. 2001;108:E20

19. Schwartz GR,Wright SW, Fein JA, Sugarman J, Pasternack J, Salhanick S.Pediatric cervical spine injury sustained in falls from low heights. AnnEmerg Med. 1997;30(3):249Y252.

20. Garton HJ, Hammer MR. Detection of pediatric cervical spine injury.Neurosurgery. 2008;62(3):700Y708.

21. Maxwell MJ, Jardine AD. Pediatric cervical spine injury but NEXUSnegative. Emerg Med J. 2007;24(9):676.

22. Hernandez J, Chupik C, Swischuk LE. Cervical spine trauma in childrenunder 5 years: productivity of CT. Emerg Radiol. 2004;10:176Y178.

23. Brockmeyer DL, Ragel BT, Kestle JR. The pediatric cervical spineinstability study. A pilot study assessing the prognostic value of four

imaging modalities in clearing the cervical spine for children with severetraumatic injuries. Childs Nerv Syst. 2012;28:699Y705.

24. Chung S,Mikrogianakis A,Wales PW, Dirks P, Shroff M, Singhal A, GrantV, Hancock BJ, Creery D, Atkinson J, et al. Trauma Association of CanadaPediatric Subcommittee National Pediatric Cervical Spine EvaluationPathway: consensus guidelines. J Trauma. 2011;70:873Y874.

25. National Institute for Health and Clinical Excellence Clinical Guidelines.Head injury: triage, assessment, investigation and early management ofhead injury in infants, children and adult. 2014;1Y63. Available at: http://www.nice.org.uk/nicemedia/live/14355/66330/66330.pdf. Accessed May2, 2014.

26. Kulaylat AN, Tice JG, Levin M, Kunselman AR, Methratta ST, Cilley RE.Reduction of radiation exposure in pediatric patients with trauma: cephalicstabilization improves adequacy of lateral cervical spine radiographs.J Pediatr Surg. 2012;47:984Y990.

27. Sun R, Skeete D, Wetjen K, Lilienthal M, Liao J, Madsen M, Lancaster G,Shilyansky J, Choi K. A pediatric cervical spine clearance protocol toreduce radiation exposure in children. J Surg Res. 2013;183:341Y346.




pediatric c-spine clearance journal club · instability, and immobilization with a cervical collar....

Documents