A previously healthy eleven-year-old girl presented to our clinic with a two- month history of neck pain and torticollis without any traumatic history or prodromal symptoms (e.g., fever). She had torticollis with a tilt of the head to the right (Fig. 1-a). Physical examination revealed restricted neck motion; lateral cervical spine bending was limited to 30° to the right and 0° to the left, and rotation was limited to 30° to the right and 40° to the left. She had no excessive ligamentous laxity. Neurologic examination revealed no abnormalities. Computed tomography (CT) showed 12° rotation to the left of C1 on C2, and 32° rotation to the right of the occipital bone on the atlas (C1) (Fig. 2-a). The diagnosis of concomitant AARF and AORF was made after CT scanning. Lateral dynamic radiographs of the cervical spine showed neither atlantoaxial nor atlanto-occipital instability (Figs. 3-a, 3-b, and 3-c).
The patient was admitted to the hospital because of the neck pain and torticollis, and she was treated with indirect traction with weight in the range of 2 to 8 kg for four weeks. Although the rotation angle at C1-C2 decreased from 12° to 7° after treatment, the rotation angle at Oc-C1 remained at 32° (Fig. 2-b). Thus, after completion of head halter cervical traction (Glisson traction) for four weeks, we treated her with halo-vest fixation for an additional four weeks. After halo-vest fixation, the rotation angle at C1-C2 improved from 7° to 2°. However, the rotation angle at Oc-C1 remained unchanged at 32° (Fig. 2-c). In spite of the residual AORF, the patient showed neither postural abnormality nor restriction of neck motion, and the neck pain disappeared (Fig. 1-b). She had no limitations in activities of daily living.
To evaluate the exact osseous morphologic lesions in detail, the atlanto-occipital articulation was scanned in the supine position with use of a three-dimensional (3D) CT (GE LightSpeed Ultra 16: scan time 0.5 s, slice thickness 1.25 mm, 80 mA, 120 kV). Data were saved in Digital Imaging and Communications in Medicine (DICOM) format and sent to a computer (Dell Precision Workstation 650, 266 MHz/2G; Dell, Round Rock, Texas). Contours of the occipital bone, atlas, and axis were semiautomatically segmented on the computer; 3D surface models were constructed based on 3D surface generation of bone cortex6 with use of a visualization toolkit-based computer program (VTK; Kitware, Clifton Park, New York). We completed surface models of the bones, which visualized 3D morphology of the craniocervical junction, by deleting the bone marrow data (Fig. 4-b). The 3D model of C1 showed protrusion of the left articular fovea and left deviated posterior tubercle (Fig. 4-a), and the 3D model of the occipital bone demonstrated protrusion of the right portion, posterior to the foramen magnum (Fig. 4-c). These findings suggested that AORF might have developed quite some time ago, for whatever reason (congenital or acquired), and that remodeling of the occipital bone and C1 was completed in the fixed position.
At five years after treatment, the patient was doing well and showed no evidence of recurrent torticollis or neck pain. A CT scan revealed that AORF remained unchanged in spite of a normalization of the atlantoaxial joints.
AARF has long been a well-known pathologic condition, having been reported since 18305. However, concomitant AARF and AORF resulting from any cause are extremely rare. To the best of our knowledge, only a few cases of AORF associated with AARF have been reported in the literature7,8. AORF is considered by some authors to be a compensatory counter-rotation mechanism in the presence of AARF9,10. Furthermore, there are several possible etiologies for AORF, including trauma, congenital anomaly, ligamentous laxity in Down syndrome, spinal tumor of the occipital condyle or the atlas, and rheumatoid arthritis11,12. According to earlier reports, AORF related to Down syndrome, rheumatoid arthritis, or tumor of the atlas is always associated with AARF12. AORF is a highly unstable condition resulting from osteoligamentous disruption between the occipital bone and C2. However, in our patient, lateral dynamic radiographs of the cervical spine showed neither atlantoaxial nor atlanto-occipital instability. Although the rotation angle at Oc-C1 did not change from 32° after a sequence of conservative treatments, the patient showed neither postural abnormality nor restriction of neck motion, and the neck pain disappeared.
Furthermore, 3D models of the occipital bone and C1 showed remodeling of these bones in the fixed position, suggesting that the patient had subclinical chronic AORF. Subclinical chronic AORF with symptomatic AARF, as in our patient, is quite rare. There likely was a permanent AORF, and the axial views really only adequately showed that the joint malalignment was chronic and, we believe, probably present since birth. The level was truly fixed and probably was not previously detected because compensation for the deformity occurred through the atlantoaxial level. The onset of symptoms caused the painful rotator problem affecting the atlantoaxial joints and did not permit compensation for the adjacent fixation deformity at the atlanto-occipital level.
We clarified the anatomic morphology in our patient by modeling each facet of the complex condition in detail with use of 3D CT studies. Because of bone overlap, particularly between the occiput and the atlas, rotation disorders and bone protrusion are difficult to evaluate with radiography, conventional CT, and magnetic resonance imaging13. The use of 3D CT has an important advantage because it can evaluate the anatomical morphology in detail, including the surface of the articulation. The occipital bone, atlas, and axis can be visualized separately with the 3D CT system. Our patient had bone protrusion between the occiput and the atlas, deformity of the occipital condyle, and asymmetric formation of the posterior arch of the atlas. It is possible that the protrusions noted on the scan were posttraumatic exostosis or response to synovial proliferation related to the rotatory fixation. That cluster of osseous abnormalities is presumed to be one of the causes of AARF.
There is some controversy regarding the best treatment for AORF and AARF. Most reports about patients with unstable AORF note that the patient underwent surgical stabilization, such as Oc-C2 fusion. However, craniocervical stabilization during childhood can potentially limit growth and cause secondary deformity14. Up to 50% of cervical rotation is lost after Oc-C2 fusion11. In spite of residual AORF, our patient showed neither postural abnormality nor restricted neck motion. In addition, the neck pain disappeared and she had no limitations in activities of daily living. Moreover, remodeling of the occipital bone and C1 in the fixed position suggested that AORF in our patient might have been subclinical and chronic. Thus, she was regarded as having adapted to this pathologic condition. Under these circumstances, the AORF could not be changed, so treatment of the AARF was the focus. For these reasons, we did not recommend surgical correction and fusion, especially because the symptomatic AARF had subsided. Although our patient was doing well five years after conservative treatment, with no evidence of recurrent torticollis or neck pain, additional long-term follow-up is needed.
Note: Katharine O’Moore-Klopf, ELS, of East Setauket, New York, performed English-language editing of this article before its final acceptance for publication.