The os odontoideum is a separate bone with smooth, rounded cortical margins, usually about half the size of the odontoid process. The gap between the os odontoideum and the second cervical vertebra (the axis) is wide and lies above the level of its superior facet1.
Whether the origin of an os odontoideum is developmental or posttraumatic remains a matter of debate2-10. In 1980, Fielding et al. reported on nine patients who subsequently had development of an os odontoideum despite radiographic documentation of a normal odontoid process immediately after trauma1. The evidence suggests that an unrecognized fracture and possible subsequent osteonecrosis might be the cause of os odontoideum.
The odontoid process begins to ossify between the first and fifth months of prenatal life. At birth, the process is separated from the body of the axis by a wide cartilaginous plate1. This radiolucent line is regularly seen in children younger than three years of age and usually is absent by six to seven years of age11. The apex of the odontoid process develops from a separate ossification center, the ossiculum terminale, which usually appears by three to six years of age and continues to enlarge, fusing with the body of the odontoid process by twelve years of age1. These synchondroses are potential weak regions that are susceptible to traumatic disruption (Fig. 1).
To our knowledge, the present case report offers the first direct evidence that an os odontoideum can arise as a sequence of posttraumatic displacement of the ossiculum terminale, as determined with use of serial high-resolution computed tomography. The family of our patient was informed that the case report would be submitted for publication, and they consented.
A five-year-old boy fell from a kindergarten slide and sustained a blow to the buttocks. Six days after the injury, he was seen in the emergency department of a local hospital because of persistent neck pain and torticollis. Initial anteroposterior and lateral radiographs of the cervical spine revealed an atlantoaxial interval of 4 mm but no obvious fracture (Fig. 2). Atlantoaxial rotatory subluxation was diagnosed on the basis of computed tomography, and the patient was transferred to our department.
The patient had not experienced a loss of consciousness and had no scalp lacerations or bruising. Neurological examination revealed normal findings. Atlantoaxial rotatory subluxation was not confirmed with computed tomography in our department, but an oblique line was identified at the tip of the odontoid process (Fig. 3). Magnetic resonance imaging demonstrated neither ligamentous nor osseous injury, except for a hematoma anterior to the dens. Three-dimensional computed tomography revealed that the cartilaginous portion between the ossiculum terminale and the odontoid process was open on the left side (Fig. 4). Given these imaging findings, epiphyseal separation of the odontoid process was diagnosed and a Philadelphia neck collar was applied. Two weeks after the application of the neck collar, the patient experienced a minor fall and a three-dimensional computed tomography scan revealed complete separation of the ossiculum terminale (Fig. 5). Halo-brace immobilization was then applied with the neck held in a position of relative extension with the patient under general anesthesia. As complete reduction of the ossiculum terminale was not achieved (Fig. 6), we attempted to reposition the ossiculum terminale by reducing the atlantoaxial interval under an image intensifier one month after brace application. Even after halo-brace immobilization for five months, a three-dimensional computed tomography scan revealed that the separation between the ossiculum terminale and the odontoid had increased (Fig. 7). We discussed with the parents the options of upper cervical fusion or close observation, and they chose close observation. At the time of the latest follow-up, more than six years after the time of the injury, the patient was eleven years old and had normal findings on neurological and clinical orthopaedic examination, although some increased motion at C1-C2 was evident on flexion and extension lateral radiographs (Figs. 8-A and 8-B). A three-dimensional computed tomography scan revealed the formation of an os odontoideum with resorption of the proximal portion of the odontoid process, leaving the ossiculum terminale to grow into an os odontoideum (Fig. 9). Follow-up continues, and atlantoaxial arthrodesis may be needed if the C1-C2 instability increases.