Chapter 7, Craniocervical Developmental Anatomy and Its Implications (continued)

Assimilation of the Atlas and Klippel-Feil Syndrome

This results from failure of segmentation between the fourth occipital sclerotome and the first spinal sclerotome (11, 13, 38). It may be unilateral, segmental, focal, or bilateral. In most instances, it occurs in conjunction with abnormalities such as the Klippel-Feil syndrome. Basilar invagination is a secondary phenomenon. The finding of atlas assimilation was present in more than 500 individuals who were evaluated for craniovertebral abnormalities (29). A hindbrain herniation occurred in 38 to 40%, caused by reduced posterior fossa volume. Segmentation failures of the second and third cervical vertebra, in association with atlas assimilation, leads to an excessive load on the atlantoaxial motion segment, which subsequently becomes unstable (Fig. 7.6A–C). Initially, a reducible atlantoaxial instability is present. The sequence of events is that there is pannus formation around the odontoid process with the reducible dislocation. However, as the child grows, there is grooving behind the occipital condyles caused by the upward migration of the axis. This is followed by a reducible basilar invagination up to approximately the age of 14 to 15 years. As these children age, the lesion becomes an irreducible basilar invagination. During the phase of reducibility and partial reducibility, a prolific granulation tissue mass crowns the odontoid process in an attempt to reduce the excursion. This, in turn, compounds the compression of the cervicomedullary junction.

In the irreducible basilar invagination, there is an associated horizontally oriented clivus, and the abnormal grooving that occurs behind the occipital condyles also pushes up the cranial base, resulting in platybasia and a short horizontal clivus (Fig. 7.7A–C). This leads to complete irreducibility of the lesion (29). Thus, a child who comes for evaluation between 4 and 16 years of age has a better chance of having a reducible atlantoaxial dislocation or a reducible basilar invagination than a fully grown adult. The upward migration of the cranial base and the reduction of the vertical height of the posterior fossa leads to an acquired hindbrain herniation syndrome (Fig. 7.8A–C). Hence, an operative procedure that relies on posterior decompression without addressing the potential for instability below the age of 20 can lead to subsequent unfortunate results.

Torticollis has been a presenting symptom in children with unilateral atlas assimilation (26). This is critical with the head manipulation that occurs during general anesthesia for patients who require placement of drainage tubes in the tympanic membrane and for adenoidectomy. In this situation, the trunk seems to stay in one position while the head is rotated 90 degrees to either side. As a result, the patients present with a rotary dislocation of the atlas on the axis. It thus behooves the treating physician to be aware of these problems, especially in children with the Klippel-Feil syndrome.

The Klippel-Feil syndrome has a classic triad of short neck, webbed neck, and a low hairline with limitation of neck motion. In this syndrome, deafness, high arch palate, facial palsies, and cardiovascular abnormalities are common. Abnormal rib fusions and scoliosis are observed, and 30% of individuals have genitourinary tract abnormalities.

Thus, the treatment of the reducible atlantoaxial dislocation or the reducible basilar invagination is stabilization; and, should a hindbrain herniation be present, a posterior fossa decompression should be made with the fusion. If, on the other hand, irreducible basilar invagination is present, ventral decompression is warranted. Unfortunately, if a posterior procedure for decompression is made in the face of ventral compression, 30% of individuals would have an unfavorable outcome (25). This is because the ventral abnormality acts as a “peg” impinging on the cervicomedullary or pontomedullary junction when the patient is positioned for the prone procedure. Additionally, fusion in a flexed position compromises the ability to perform a satisfactory ventral decompression. It is important to reiterate that the ability to reduce the invagination is age related. The presence of syringohydromyelia with hindbrain herniation and basilar invagination should not sway the neurosurgeon to perform a posterior operative procedure alone. In most instances, the syringohydromyelia disappears once the ventral abnormality has been corrected.

Os Odontoideum

The ossiculum odontoideum is an independent bone in the place of the dens and is cranial to the axis (38). It should not be considered as an isolated dens but exists apart from a small hypoplastic dens. Radiographically, the os has smooth borders and a small dens is always present. It is located in the position of the odontoid process near the basiocciput, where it may fuse with the clivus. The gap between the axis and the free ossicle usually extends above the level of the superior facets of the axis, thus making this an acquired abnormality rather than a congenital abnormality (Fig. 7.9A–C). The entire complex leads to an incompetence of the cruciate ligament and, subsequently, to atlantoaxial instability (25).

In our series, the evidence for os odontoideum has pointed to trauma between the ages of 1 and 4 years in cases where there is a previously recognized normal odontoid process (29). At times, os odontoideum may be associated with an unrecognized fracture in children less than 5 years of age with previously normal odontoid structure, as observed in a large number of patients in our series (8, 25). Os odontoideum may also be associated with nontraumatic situations with ligamentous laxity, such as Down’s syndrome and Morquio’s syndrome (34). In our series, symptomatic patients were found to have instability in all planes. The biomechanics are complex. It is different for each individual, in that, in some situations, the flexed position may be the best to relieve compression of the cervicomedullary junction, whereas in others, it is the extended position. Thus, the biomechanics must be carefully studied.

Irreducible dislocation was found to be caused by 1) pannus and 2) a cruciate ligament that slipped behind the os odontoideum and in front of the superior portion of the dens. Thus, both the dens and the axis body create the compressive mechanism on the ventral cervicomedullary junction; not the ossicle alone. In severe chronic dislocations, the os may become fixed with severe basilar invagination. At its worst, os odontoideum has significant implications regarding compression of the cervicomedullary junction (Fig. 7.10A–C). I strongly think that all patients with recognizable instability at the craniocervical junction and associated os odontoideum should undergo stabilization. We presently use lateral mass screw/rod fixation or transarticular screw fixation between C2 and C1. In some circumstances, an occipitocervical arthrodesis may be necessary. If a fixed irreducible abnormality is found, a decompression is first performed in the manner in which compression occurred.

Atlas Abnormalities

Failure of development and failure of segmentation of the atlas result in abnormal articulation between the clivus, the atlas, and the odontoid process. Variations may consist of partial absence of the posterior arch of the atlas and a bifid atlas anteriorly or posteriorly. A bifid anterior or posterior atlas results in the two halves of the atlas vertebrae acting like a complex Jefferson fracture with lateral displacement (2, 3, 6, 17, 20, 23, 31). This condition should be considered grossly pathological if it is present beyond the age of 3 years and should be addressed (Fig. 7.11A and B).

I have reviewed 160 normal CT scans of the craniovertebral junction in subjects between the ages of birth and 4 years of age. Our conclusion was that the atlas should be a complete ring by 3 years of age. In cases where the atlas persists as being separate with abnormal dynamics, an operative intervention or fusion should be performed. Neurological deficit was observed in 16 infants with a bifid anterior and posterior arch of the atlas or with absent anterior and posterior arches, except for preservation of the lateral atlantal masses. Placement in a custom-built cervical collar and bracing through 3 years of age has allowed for reformation of the anterior atlas arch and stabilization of the craniocervical junction in 60% of children. It is thought that continued motion in an abnormal situation will prevent the formation of the absent segments and, thus, lead to further neurological deficit.

A persistent bifid anterior and posterior arch of the atlas beyond the age of 3 to 4 years is observed in skeletal dysplasias, Down’s syndrome, Goldenhar’s syndrome, Conradi syndrome, and atlas assimilation. In our series, the presentation was torticollis and plagiocephaly. Neurological dysfunction presented as paresis and apnea and failure to thrive.

The ideal form of imaging is 3-D CT and magnetic resonance imaging (37). As previously mentioned, the treatment in our series consisted of bracing until 4 years of age with repeat 3-D CT on a yearly basis (Fig. 7.12). If instability was still present beyond age 4 years, an occipitocervical bony arthrodesis was accomplished.

Axis Spondylolysis

The rare situation of axis spondylolysis has only recently been recognized, because of the improved neurodiagnostic imaging of 3-D CT and the improved motion dynamics with magnetic resonance imaging. A persistent pars defect led to abnormal bone formation in an older individual. In a very young patient, I have observed significant kyphosis and canal compromise at the affected pars defect at the axis vertebrae. The stability of the craniocervical unit is dependent on the geometry of the articular surfaces of the occipital condyles, and the lateral atlantal and axial masses. The integrity is also maintained by the ligamentous complex in the anterior and the middle column. However, the dorsal column is unstable, with a freely mobile segment of bone beyond the pars that may cause hypertrophy at this junction, causing lateral as well as dorsal compression. Decompression and stabilization has been performed in the younger individual.

In an older patient, a pincers-like bony ingrowth circles the spinal cord behind the axis vertebra (Fig. 7.13A–D). Thus, decompression for axis spondylolysis was mandated by an anterior procedure going through the entire vertebral body and resecting the area of bony compression at the level of the pars defect. This complex situation was encountered in the patient shown in Figure 7.13.


A large database of 4700 patients has provided for the understanding of the natural history of many entities. This database has allowed treatment protocols to be established that have stood the test of time. In addition, the previous presentation in this manuscript confirms the embryology of the craniocervical junction via a clinicopathological correlation. It is important to obtain newer imaging techniques to unravel the complexities of the craniocervical junction.

Previous Page | Page 41 Next Page