Odontoid fractures in children younger than seven years of age represent a
disruption through the cartilage plate (synchondrosis) that connects the
odontoid to the body and neural arches of the
axis1,2.
Unrecognized and untreated fractures of the
odontoid3-5
and acute ligamentous
injuries6 in young
children may compromise the vascularity of the process, causing the
development of an os odontoideum. Because the resulting atlantoaxial
instability may be asymptomatic or produce minimal complaints, the diagnosis
of an odontoid abnormality can be delayed for several
years3-6.
We present the cases of two children who sustained a fracture of the
odontoid at an undetermined time prior to diagnosis. Despite the delayed
diagnosis, both patients were managed successfully nonoperatively. The parents
were informed that data concerning these cases would be submitted for
publication.
Case 1
In April 1993, a very active four-year and nine-month-old boy was seen at a
hospital in another state because of headache and neck stiffness. No injury
was reported at that time. He underwent magnetic resonance imaging of the
brain and serum testing for C-reactive protein. The magnetic resonance imaging
scan revealed normal findings, and the C-reactive protein level was within
normal limits.
Two weeks later, he was seen in the emergency department of a local
hospital with a very stiff neck. The neurological examination revealed normal
findings, as did the radiographic skeletal survey. Cutaneous signs of child
abuse were absent. The mother reported that the child had had development of
acute neck pain and stiffness two to three weeks earlier. However, for the
past year, the boy had experienced short-lasting episodes of intermittent neck
pain that spontaneously resolved.
Cervical spine radiographs demonstrated a fracture of the odontoid, with 8
mm of anterior displacement and a severe local kyphotic deformity
(Fig. 1-A). Retropharyngeal
soft-tissue swelling measured 14 mm at the lower end plate of C2. Computerized
tomographic scout images revealed that the fracture reduced with neck
extension (Fig. 1-B).
The child was transferred to our hospital. With the patient under general
anesthesia, a halo connected to a plaster body jacket was applied. Six halo
fixation pins were used, with a reduced insertion torque of 4 in-lb. The
kyphotic angulation at the fracture site was improved; however, 3 mm of
anterior displacement persisted (Fig.
1-C). One month later, one of the metal supports that connected
the halo to the cast cracked, and it was replaced. The fracture reduction
remained unchanged during the time of halo wear despite a high level of
activity and a severe fall from a bicycle. The pin sites became infected three
months after the application. The halo cast was removed, and a plaster body
jacket with a neck mold was applied for six more weeks. Flexion and extension
radiographs that were made four and one-half months after the initial halo
treatment demonstrated fracture-healing. Five years later, a cervical spine
radiograph demonstrated remodeling at the site of the odontoid fracture
(Fig. 1-D).
At the age of seventeen years, the boy reportedly experienced occasional
neck pain that did not require medication and continued to be very physically
active.
Case 2
In November 1996, a three-year and three-month-old girl was brought to the
emergency department after the sudden development of severe neck pain, with no
known prior injury. She had stepped out of the family car and walked into her
home, complaining of neck pain. She cried and refused to move her head and
neck. She kept her head slightly tilted upward and, at times, held it with
both hands.
Five months previously, the child and her five siblings had been removed
from their biological parents' home because of reported abuse. The child had
developmental delays, and she could not speak.
The foster parents noticed that she could not chew solid foods or fully
open her mouth. Occasionally, she held her neck in extension. She did not
appear to be in pain until the day of presentation to the hospital. The
neurological examination revealed normal findings, and no cutaneous signs
consistent with recent abuse were observed.
Radiographs of the cervical spine showed a fracture of the odontoid, with 5
mm of anterior displacement (Fig.
2-A). The fracture reduced with extension of the neck
(Fig. 2-B). The retropharyngeal
space measured 5 mm at the lower end plate of C2.
The child was then transferred to our hospital for care. The child was
given general anesthesia for the application of the halo, but the
anesthesiologist was unable to open the patient's mouth for intubation. He
reduced a dislocated temporomandibular joint dislocation with downward and
then forward pressure, and then he could open the patient's mouth. The halo
was applied and connected to a plaster body jacket cast with the patient's
neck held in relative extension (Fig.
2-C). Eight halo fixation pins were used, with a reduced insertion
torque of 4 in-lb. Postoperatively, the patient was evaluated by an oral and
maxillofacial surgeon but required no further treatment for the
temporomandibular joint dislocation. She was able to smile, chew her food, and
open her mouth widely. The halo was removed after eleven weeks, at which time
flexion and extension lateral cervical spine radiographs demonstrated no
motion at the fracture site. A soft collar was worn for two weeks for
comfort.
One year later, the patient had no complaints involving either the neck or
the jaw. Nine years after the fracture, she had no complaints and the odontoid
process appeared normal (Fig.
2-D). She could speak quite well but had a learning
disability.
In children, the synchondrosis between the odontoid and the body of the
axis is particularly vulnerable to injury, with the average age at the time of
the injury being three
years7. These
fractures can result from high-energy
trauma1,2,7-11
as well as from more trivial
injuries7,12.
The treatment of choice for acute odontoid fractures is conservative
management, with reduction by extension of the neck and immobilization in
either a Minerva cast, a halo cast, or a
halo-brace1,2,7-11.
We found only one previous report of an established nonunion of the
odontoid in a child, a five-year-old
boy11. The
diagnosis had been delayed for four months after the time of injury. The
patient underwent a primary posterior C1-C2 fusion. Postoperatively, a deep
infection developed, the implant was removed, and the graft resorbed.
Nonetheless, the fracture healed after the patient had been placed in a
Minerva cast for three months.
Seimon12 reported
on a child who had an untreated, undisplaced fracture of the odontoid process.
The odontoid fracture gap progressively widened on segmented radiographs,
similar to changes seen in association with a nonunion of the carpal
scaphoid.
Our two patients had several features in common. Both had unstable family
conditions and an absence of acknowledged injury, leading to suspicions of
child abuse and neglect. In the case of one patient (Case 2), child abuse was
confirmed to have occurred five months before the time of presentation for the
treatment of neck pain. In that case, the temporomandibular joint dislocation
was missed on the initial evaluation because the patient's reported inability
to speak, to chew solid foods, and to open her mouth widely was thought by the
surgeon to be part of her developmental delays. Neither of these children had
lacerations, scars, or bruises consistent with recent
abuse13. In both
cases, the caretakers reported that the children had had minimal complaints
for several months before the onset of acute symptoms. It is possible that
these children could have had a nonunion of the odontoid and then experienced
a trivial injury that acutely increased odontoid displacement.
There appears to be a critical age at which trauma to the upper cervical
spine may result in an os odontoideum. There is no documented case in which an
acquired os odontoideum has been reported following a neck injury in a child
older than three years of
age5.
The fractures in our two patients reduced easily with neck extension as a
hinge of periosteum likely existed
anteriorly10. After
reduction, the halo
cast2,7,8,11
was used for immobilization. Because of the decreased thickness of a child's
skull, six halo fixation pins were used in one patient (Case 1) and eight pins
were used in the other (Case 2), with a reduced insertion torque of 4
in-lb14-16
instead of the four pins at 8 in-lb as used in adults. At the time of
treatment, plastic vests for children were not commercially available, so a
plaster body jacket was used instead. The fractures appeared to be healed in
eight to sixteen weeks, the time expected for acute odontoid fractures.
We recommend a trial of conservative treatment for fractures of the
odontoid of undetermined time of occurrence in children younger than seven
years of age, when the synchondrosis is still open. ?