After institutional review board approval, our spine database was used to
identify all children diagnosed with occipitalization since August 1986. Of
the thirty-three patients identified, thirty were included in the study
because of the availability of adequate patient information and imaging
studies. The minimum radiographic inclusion requirements were anteroposterior
and lateral radiographs of the cervical spine, dynamic lateral flexion and
extension cervical spine radiographs, and at least one computed tomography or
magnetic resonance imaging study of the cervical spine. The clinical data
collected included patient age and gender, the reason for clinical
presentation, and any associated syndrome or additional disorder. The imaging
data collected included the morphologic pattern of occipitalization (discussed
below) as well as the presence of atlantoaxial or subaxial instability,
associated subaxial fusion, basilar invagination, Chiari malformation, spinal
canal encroachment, an open posterior arch of the atlas (rachischisis), os
odontoideum, or intraspinal or central nervous system abnormalities such as
syringomyelia, diastematomyelia, or hydrocephalus. Clinical management and
outcomes of treatment were not considered in this study.
Four orthopaedic surgeons and one pediatric neuroradiologist reviewed all
of the imaging studies. The radiographic techniques and definitions are
described below.
Occipitalization
On cervical spine radiographs, occipitalization was suspected when the
atlas and occiput showed osseous continuity or moved as a functional unit
(i.e., there was no evidence of motion between the two) on lateral flexion and
extension radiographs of the cervical
spine8,16.
The morphologic pattern of occipitalization was identified on
two-dimensional sagittal and coronal reformatted computed tomography
reconstructions (0.5-mm-thick cuts in a plane truly perpendicular or parallel
to the C1 arch to avoid any errors resulting from parallax or angulation of
the spine due to positioning). Occipitalization was identified as an osseous
continuity between the occiput and atlas on these reconstructed images.
Coronal and sagittal magnetic resonance images of the cervical spine were
utilized to identify the morphologic pattern of occipitalization in seven
patients who had not had computed tomography scans. On T2-weighted images, the
bone marrow in children usually appears bright whereas the bone cortex appears
dark, as a result of a relative paucity of protons in the bone
cortex17,18.
Emery et al.19
confirmed tarsal coalition (osseous fusion) on T2-weighted magnetic resonance
imaging scans of the foot by observing continuity of bone marrow across the
tarsal joints. In our study, continuity of bone marrow or cortex between the
skull and atlas on the T2-weighted magnetic resonance images was viewed as
evidence of occipitalization.
Classification
Occipitalization was classified into zones on the basis of whether the
failure of segmentation was in the anterior arch, the lateral masses, or the
posterior arch of the atlas. The anterior arch was defined as the part of the
atlas extending anteriorly from the lateral masses. The lateral masses were
defined as the bilateral osseous prominences of the atlas, including the
transverse processes. The posterior arch was defined as the part of the atlas
extending posteriorly from the lateral masses
(Fig. 1). Zone-1
occipitalization was fusion involving the anterior arch of the atlas
(Fig. 2, A and
E). Zone-2 occipitalization was fusion involving the
lateral masses of the atlas (Fig. 2,
B and F). Zone-3 occipitalization was fusion
involving the posterior arch of the atlas
(Fig. 2, C and
G). Fusions involving more than one zone of the atlas
were classified as a combination of zones
(Fig. 2, D). Within
each zone, occipitalization could be partial or complete and unilateral or
bilateral.
Atlantoaxial and Subaxial Instability
Atlantoaxial instability was defined as an atlantodental interval of >4
mm as observed on lateral flexion and extension radiographs of the cervical
spine16,20-22.
The atlantodental interval was measured from the anterior surface of the
odontoid to the posterior cortex of the anterior arch of the
atlas16,22,23.
In patients who lacked an intact odontoid (had os odontoideum), the
atlantodental interval was measured from the anterosuperior part of the body
of the axis, where the odontoid should normally be attached to the posterior
cortex of the anterior arch of the
atlas24.
Subaxial instability was defined as =2 mm of translation or >11°
of angulation between adjacent cervical vertebrae as noted on lateral flexion
and extension radiographs (Fig. 3,
A and
B)25,26.
Vertebral Fusion
Vertebral fusions may be complete block fusions or partial fusions
involving either the posterior elements of the vertebrae or both the anterior
and the posterior elements (Fig. 3,
A and B). With fusion of the posterior elements
only, the normal gap between adjacent posterior elements is eliminated, with
the spinous processes moving as a functional unit. Block fusions are evidenced
by fusion of both the posterior elements and the vertebral bodies of adjacent
vertebrae as noted on the lateral radiographs. Because of the unique
embryology and somatogenesis of the upper cervical spine, we define upper
cervical spine fusions as those occurring between the occiput and the C2-C3
disc space and lower cervical spine fusions as those occurring caudad to the
C2-C3 disc space down to the C7
vertebra27,28.
For the purpose of this study, Klippel-Feil
syndrome29 was
defined on the basis of both radiographic and clinical
characteristics—i.e., it was defined as vertebral fusion in the lower
cervical spine associated with any two clinical criteria, which included a
short neck, a low posterior hairline, or a limited range of motion of the
neck.
Basilar Invagination
Basilar invagination has historically been measured on the basis of the
plain radiographic criteria described by McRae and
Barnum8,
Chamberlain30,
Fischgold and
Metzger31, and
McGregor32.
However, reference points near the base of the skull are not always easy to
identify on radiographs and hence computed tomography and magnetic resonance
imaging scans were used to diagnose basilar invagination in our study. A
sagittal reformatted computed tomography scan showing the tip of the odontoid
at or above the level of the foramen magnum was accepted as evidence of
basilar invagination (Fig. 2,
D)33.
On magnetic resonance imaging, a midsagittal section demonstrating the tip of
the odontoid at or above the level of the foramen magnum or a short and
horizontal clivus with the margin of the foramen magnum above the floor of the
posterior fossa was accepted as evidence of basilar invagination
(Fig. 4,
A)1.
Spinal Canal Encroachment
We defined spinal canal encroachment as a reduction (to =13 mm) in the
space available for the spinal cord. Spinal canal encroachment may be due to
static or dynamic factors, or both. Static encroachment occurs as the result
of focal osseous narrowing of the spinal canal (stenosis), the presence of a
space-occupying mass (such as an intrusion of the cerebellar tonsils with a
Chiari malformation), or protrusion of the odontoid into the foramen magnum as
with basilar invagination. Dynamic encroachment occurs with segmental cervical
spine instability as translational encroachment on the spinal cord takes place
with neck movement.
The space available for the spinal cord is measured from the posterior
aspect of the odontoid to the anterior aspect of the posterior arch of the
atlas22,34.
Occipitalization can obscure the osseous landmarks on radiographs, so these
measurements were made on magnetic resonance imaging or computed tomography
scans (Fig. 4, B).
Open Posterior Arch of the Atlas (Rachischisis)
An open posterior arch of the atlas was confirmed on axial computed
tomography or magnetic resonance images. A cartilaginous cleft and an open
synchondrosis between the osseous posterior arches of the atlas is a normal
finding in children up to four years of
age11,35;
therefore, a true defect in the posterior arch was diagnosed only in patients
who were more than four years old.
Os Odontoideum
Os odontoideum is defined as an anomaly of the odontoid process of the
second cervical vertebra when the odontoid is replaced by a well-circumscribed
ossicle bordered by a circumferential cortical margin. The ossicle is the
presumed remnant of the odontoid and can be situated in either the normal
position (orthotopic) or an abnormal one
(dystopic)24,36.
In this study, when this anomaly was suspected from the plain radiograph its
presence was confirmed by computed tomography or magnetic resonance
imaging.
There were sufficient data to include thirty of the thirty-three identified
cases of occipitalization in the study. Three patients were excluded as we
were unable to locate their imaging studies (computed tomography or magnetic
resonance imaging scan) for review. There were twenty-four boys and six girls.
The mean age at the time of the diagnosis was 6.5 years (range, 0.5 to fifteen
years). All patients included in the study had anteroposterior, lateral, and
lateral flexion-extension radiographs of the cervical spine. Twenty-eight
patients had a magnetic resonance imaging scan of the cervical spine (with
fifteen of them having a flexion-extension magnetic resonance imaging scan),
and twenty-three patients had a computed tomography scan of the cervical
spine. Twenty-one patients had both a magnetic resonance imaging and a
computed tomography scan of the cervical spine, and all patients had at least
one of these imaging studies.
Seventeen (57%) of the thirty patients had a syndrome or a presumed
syndrome. Six patients had Klippel-Feil syndrome, which was associated with
VATER syndrome in two and with Russell-Silver syndrome in one. Four patients
had a 22q11.2 deletion syndrome. One patient had Goldenhar syndrome, and one
had multiple hereditary exostoses. Five patients had other visceral and
appendicular anomalies, but these patients either refused or did not have
additional diagnostic testing.
The initial clinical presentation varied. Eleven patients (37%) presented
with a painful neck, with six having had an insidious onset of the pain and
five having the pain following a minor neck injury. Occipitalization was
identified in twelve syndromic patients as part of a radiographic skeletal
survey to rule out other anomalies. Occipitalization was an incidental finding
in two patients. Five patients presented with myelopathy. Physical examination
revealed restricted neck motion in twelve patients, with four of them having
torticollis.
The five patients with myelopathy had an upper motor neuron pattern of
weakness, which was mild in three and more pronounced in two. Of the two
patients with pronounced weakness, one was unable to walk and the other was
able to walk with assistance. Three of the five patients had sensory
hypesthesia associated with the motor deficit. One patient, who previously had
been continent, presented with bowel and bladder incontinence at the age of
8.5 years. None of the patients presented with cranial nerve
abnormalities.
Occipitalization was initially suspected in all patients when lateral
flexion and extension radiographs of the cervical spine showed no evidence of
motion between the atlas and occiput. This was confirmed in all patients by
computed tomography and/or magnetic resonance
imaging19. The
morphologic pattern was Zone 1 in six patients (20%), Zone 2 in five (17%),
Zone 3 in four (13%), and a combination of zones in fifteen (50%). Two of the
fifteen patients who had a combination of zones of occipitalization had fusion
in all three zones, with one of the two having total (block) occipitalization.
Of the remaining patients with a combination of zones of occipitalization,
five had fusion in Zones 1 and 2; four, in Zones 1 and 3; and four, in Zones 2
and 3.
Seventeen patients (57%) had atlantoaxial instability, with a mean
translation in the atlantodental interval of 6.5 mm (range, 5 to 11 mm). Eight
of the seventeen patients with atlantoaxial instability had an associated
C2-C3 fusion. Overall, fourteen patients (47%) had a C2-C3 fusion (twelve had
a block fusion and two had a fusion of the posterior elements). Lower cervical
spine fusion (caudad to the C2-C3 disc space) was noted in six patients (20%),
and all had a block fusion. Only one patient had evidence of subaxial
instability, which consisted of C3-C4 instability between a C2-C3 block fusion
and a C4-C5-C6 block fusion (Fig. 3,
A through D).
Eleven patients (37%) had basilar invagination. Five patients had a Chiari
type-I malformation (Fig. 4,
C), three had cerebellar tonsillar ectopia, and none had
a Chiari type-II malformation. Eleven patients (37%) had spinal canal
encroachment, defined as space available for the spinal cord of 13 mm (range,
7 to 13 mm in the present study). Static encroachment was observed in four
patients; dynamic encroachment, in five; and combined encroachment, in two.
All eleven patients with spinal cord encroachment had some degree of dural
effacement, and five of those eleven patients had myelopathy on
examination.
An open posterior arch of the atlas was noted in five patients (17%), at a
mean age of 8.4 years (range, four to eleven years). Four patients with an
open posterior arch of the atlas had associated atlantoaxial instability, and
one had myelopathy. An os odontoideum was identified in two patients, and both
had atlantoaxial instability. Four patients had syringomyelia; two had it in
the cervical spine and two had it in the thoracic spine, and three of these
patients also had a Chiari type-I malformation. Two patients had
hydrocephalus, which was associated with a Chiari type-I malformation in
both.
Eleven of the fourteen patients with C2-C3 fusion and eight of the eleven
patients with basilar invagination had occipitalization in Zone 1 and/or 2.
Seven of the eleven patients with spinal canal encroachment had
occipitalization in Zone 2. None of the patients with Zone-2 occipitalization
had spinal cord anomalies. As a result of the size of the cohort, we were
unable to demonstrate a significant association between the zones of
occipitalization and the clinical and radiographic findings. However, both
C2-C3 fusion and basilar invagination were most frequently present in patients
with occipitalization in Zone 1 and/or Zone 2.
The objectives of this study were to evaluate the clinical and radiographic
findings in children with occipitalization (1) to identify morphologic
patterns of occipitalization on two-dimensional sagittal and coronal
reformatted computed tomography reconstructions and/or magnetic resonance
images; (2) to identify other anomalies associated with occipitalization and
their sequelae; and (3) to correlate the above findings with the patient's
prognosis with regard to the development of myelopathy. To the best of our
knowledge, the current study represents the largest series of children with
occipitalization.
Erbengi and
Oge15 classified
anomalies of the craniovertebral junction on an embryological basis. They
suggested that a defect in a single germ layer be termed a "minor
anomaly," while a defect in two germ layers be labeled a "major
anomaly." Occipitalization was classified as a minor anomaly with this
system, with no further attempt to delineate the pattern of fusion. We were
able to identify and categorize all cases of occipitalization in our series
with a simple anatomic system based on zones of the atlas. Zone 1 indicated
the anterior arch; Zone 2, the lateral masses; and Zone 3, the posterior
arch.
McRae and Barnum8
and McRae9 reported
on combined (adult and pediatric) cohorts of twenty-five and twenty-eight
patients with occipitalization, with findings based on radiographs,
laminograms, and myelograms. Bharucha and
Dastur13 as well as
Erbengi and Oge15
subsequently reported on their series of patients with craniovertebral
anomalies, who were examined with use of radiographs and
myelograms13 or
with use of radiographs, computed tomography, and magnetic resonance
imaging15.
Pertinent findings in patients with occipitalization in these series are
compared with our findings in Table
I.
McRae9 reported
that twenty-two (79%) of the twenty-eight patients in their study were male.
Similarly in our cohort, twenty-four (80%) of the children were male. These
findings suggest that occipitalization has a male predominance.
Craniovertebral anomalies have been noted to be associated with congenital
syndromes, skeletal dysplasias, and metabolic
conditions7,20,22,37.
In our series, seventeen patients (57%) had a syndrome or a presumed syndrome,
with six having Klippel-Feil syndrome and four having 22q11.2 deletion
syndrome.
Neck pain and stiffness have been reported to be presenting symptoms in
patients with occipitalization. Bharucha and
Dastur13 reported
that these were the presenting symptoms in sixteen of twenty-three patients
with occipitalization; ten of these sixteen patients reporting a precipitating
neck injury, the nature of which varied from a sneeze to a fall from a height
of 30 ft (9 m).
McRae9 reported neck
pain and stiffness in eleven of twenty-eight patients. In our series, eleven
patients presented with neck pain and stiffness, with trauma as a
precipitating factor in five of them. The nature of the trauma was usually a
hyperextension injury of the neck while the child was playing sports. Thus,
neck pain appears to be a relatively common presenting feature in these
patients; however, neck pain may be absent in many cases of
occipitalization.
Neurological signs and symptoms in patients with occipitalization typically
present in the third or fourth decade of
life7-9,13,15.
McRae9 reported
neurological symptoms in nineteen of twenty-eight patients, with a mean age of
onset of thirty-one years (range, seven to forty-five years). Fourteen
patients presented with motor weakness, eleven had sensory symptoms, and
thirteen had spasticity. Erbengi and
Oge15 reported
neurological symptoms in fourteen of twenty patients, with a mean age of onset
of 32.5 years (range, sixteen to fifty-seven years). In our study, which was
limited to children, five patients had neurological symptoms with a mean age
of onset of 5.5 years (range, one to 9.8 years). On the basis of the
relatively high prevalence of myelopathy in earlier
reports7-9,13,15
on cohorts that mainly included adult patients, we expect the prevalence of
myelopathy in our series to increase with time. We suggest advising families
regarding this expected natural history, stressing the need for subsequent
periodic clinical and radiographic examinations.
Historically, narrowing of the osseous spinal canal has been described as
spinal canal stenosis (the space available for the spinal cord being =13
mm)7,14,20,22.
The term "spinal canal stenosis" is inadequate to include all of
the pathologic conditions that decrease the space available for the spinal
cord, and hence we defined the concept of spinal canal encroachment.
Encroachment of the spinal canal may be due to static or dynamic factors, or
both. Static encroachment includes focal osseous narrowing of the spinal canal
(stenosis), the presence of a space-occupying mass such as an intrusion of the
cerebellar tonsils with Chiari malformation, or protrusion of the odontoid
into the foramen magnum as with basilar invagination. Dynamic encroachment
occurs with segmental instability as translational vertebral motion narrows
the space available for the spinal cord. Although some degree of dural
effacement was encountered in all eleven patients with spinal cord
encroachment, this was not considered to be a criterion for encroachment
because of the variation in the subjective description of dural effacement.
All five patients with myelopathy in our series had spinal canal encroachment.
Factors leading to encroachment, such as atlantoaxial
instability22,38,39,
spinal canal
stenosis40,41,
basilar
invagination42,43,
and Chiari
malformation43,44,
are known to cause
myelopathy22,38,40,41,44.
Overall, atlantoaxial instability was noted in seventeen patients in this
series. Eight of them had an associated C2-C3 fusion, and nine did not.
Atlantoaxial instability was observed by
McRae9 in fifteen of
twenty-eight patients, and eleven of them had an associated C2-C3 fusion.
Occipitalization associated with C2-C3 fusion appears to place additional
strain on the atlantoaxial articulation during flexion and
extension9,13,20.
Repeated excessive strain at this joint likely predisposes a patient to
atlantoaxial instability. A careful clinical and radiographic assessment of
the cervical spine is recommended for patients with occipitalization,
particularly when it is associated with C2-C3 fusion, in order to identify
atlantoaxial instability early and provide treatment to prevent
myelopathy.
As others have noted (Table
I), we observed basilar invagination (in eleven patients) and
Chiari type-I malformation (in five). Five patients had an open posterior arch
of the atlas, a finding reported by
Geipel45 in 4% of
adult autopsy specimens, and four of those five patients in our study had
associated atlantoaxial instability. Although this association was noted,
these numbers are small and it may be premature to conclude that an open
posterior arch of the atlas is likely to predispose to the development of
atlantoaxial instability. Knowing if there is an open posterior arch of the
atlas is important if surgery through a posterior approach is planned. Two of
our patients had an os odontoideum with atlantoaxial instability, a
combination that has not been previously reported, to our knowledge.
Only one patient in our series had a total (block) occipitalization of the
atlas, while half of the patients had a fusion in more than one zone. The only
findings in the current study that showed a potential association between
anomalous development and specific zones of occipitalization were a C2-C3
fusion and basilar invagination. Eleven of fourteen patients with C2-C3 fusion
and eight of eleven with basilar invagination had an occipitalization in Zone
1 and/or 2. Zone-2 occipitalization also had the highest association (63%)
with spinal canal encroachment. The associations that we noted with different
zones of occipitalization have not been previously described, to our
knowledge. Although a significant correlation was not found, such
relationships may be important in terms of patient prognosis with increasing
age.
In summary, distinguishing patterns of occipitalization is difficult
without two-dimensional sagittal and coronal reformatted computed tomography
reconstructions and/or magnetic resonance images. The morphologic pattern of
occipitalization and other congenital osseous anomalies are best visualized
with bone window settings on computed tomography. Magnetic resonance imaging
(with flexion-extension views) is recommended to rule out static or dynamic
canal encroachment, which is frequently associated with occipitalization and
includes atlantoaxial instability, canal stenosis, basilar invagination, and
Chiari malformation. Furthermore, occipitalization is associated with an
increased risk of atlantoaxial instability developing, particularly when there
is an associated C2-C3 fusion. We noted an association of basilar
invagination, canal encroachment, and C2-C3 fusion with occipitalization of
Zone 1 and/or 2, but additional studies are necessary to corroborate these
findings. Finally, occipitalization is associated with an increased risk of
myelopathy and serious neurological sequelae. Although few neurological
deficits are evident during childhood, previous reports on the natural history
of occipitalization suggest that these deficits may increase with age.
Families of children with occipitalization should be educated about potential
spinal cord compression and myelopathic problems and the need for these
patients to seek regular clinical and imaging evaluations throughout their
adult lives. ?