Analysis of surgical strategies and efficacy in the treatment of Os odontoideum with atlantoaxial dislocation

A total of 32 males and 24 females aged 49.36 ± 13.98 years were enrolled, whose course of disease was 8.9 ± 3.38 months (Table 1). Inclusion criteria included: (1) Computed tomography (CT) coronal reconstruction presented the OO, and anterior atlanto-dental interval (ADI) > 3 mm in lateral X-ray and (2) those with acute or chronic clinical manifestation as upper motor neuron injury clinical manifestation of upper motor neuron injury. Exclusion criteria included: (1) AAD caused by infection, tumor, or tuberculosis, (2) coexisting basilar invagination, and (3) history of atlantoaxial surgery or surgery in surrounding areas.

The study was in accordance with the Declaration of Helsinki and was approved by the Ethics.

Committee of The Second Affiliated Hospital of Xi’an Jiaotong University (Xi’an, China). Written informed consent was obtained from all patients.

All the operations were performed by the corresponding author. Except for type I cases, all cases were treated with post anesthesia skull traction with a maximum weight of 1/6 of body weight. Type I and II cases were treated with posterior fixation and fusion. Type III cases received posterior fixation and fusion after irreducible dislocations were converted to reducible dislocations by trans lateral mass release or transoral release. Type IV cases required transoral release for conversion into reducible dislocations before posterior fixation and fusion. Here, the posterior side of the lateral atlantoaxial joint was exposed for patients who underwent trans lateral mass release, the vessels and nerve bundles between C1 and C2 were carefully protected, The C1 and C2 lateral mass joint capsule were incised, The loosening of the joints was completed by the axial distraction of the joints with a 7-mm-wide bone inserted knife into the facet joints, The atlas was levered by the bone knife, and the axis was pushed to reduce AAD; this procedure was monitored through fluoroscopy. After lateral mass joint release, direct posterior fixation and fusion were then performed. In patients with transoral release, Initially, the patient was placed supine, The transoral approach was performed with the assistance the mouth retractor, which could expand the mouth and exposed the posterior wall of the pharynx. The posterior pharyngeal wall layer was longitudinally incised along the middle line. Then, the anterior atlas arch, the axis vertebral body, and the lateral mass joints were gradually exposed toward bilateral, cephalic, and caudal sides. The contractural muscles, the anterior longitudinal ligament, and joint capsules between the atlantoaxial joints were incised and released, and the osteophytes and scar tissues between C1 and the C2 vertebral body, and the apical and alar ligaments were cleaned. The reduction of AAD was then performed by traction and leverage. Then, turn over, take the prone position, Posterior fixation and fusion in the end was then performed. The pedicle screw is the most effective and commonly used fixation technique for the C2 vertebra, for this reason, the pedicle screw was used in most of our cases. But in the case of the high-riding vertebral artery, the pedicle was forced insertion of a screws, which easily damages the medial spinal cord and the lateral vertebral artery (VA), laminar screw was used for fixation in this case.

The patients were followed up for 12–24 months (13 months on average). The operation time, intraoperative blood loss, and perioperative complications were recorded. The neurological function was assessed with the 17-point Japanese Orthopedic Association (JOA) scoring system, which addresses upper- (4 points) and lower-extremity motor function (4 points), upper- (2 points) and lower-extremity sensory function (2 points), trunk sensory function (2 points), and bladder function (3 points). Other indicators were evaluated by imaging. In detail, X-ray or CT of the cervical spine was conducted at 3 and 6 months after operation, from which the anatomical relationship of atlantoaxial vertebra and the callus of bone graft and healing status were observed. In order to understand whether there is still spinal cord compression after surgery, some patients were reexamined by MRI.

SPSS19.0 was used for statistical analysis. JOA scores before and after operation are presented as mean ± standard deviation (\({\overline{\text{x}}}\) ± s), and one way ANOVA analysis was used to compare them P < 0.05 denoted a statistically significant difference.

In this study, type I and type II cases [n = 40 (71%)] were in the majority, suggesting that OO with AAD is mostly reducible. Type III and type IV cases [n = 16 (29%)] were often associated with rheumatoid arthritis, ankylosing spondylitis, and other diseases.

Previously, posterior atlantoaxial arthrodesis was performed with fixation of the posterior arch of atlas, and the surgical procedures included Gallie steel wire fixation of the posterior arch of atlas and the spinous process of the axis, Brooks wire fixation, and Halifax clamp fixation [12, 13]. However, an intact posterior structure is required for these methods, and the reconstruction effect of the atlantoaxial stability is unsatisfactory [14]. With the development of the pedicle screw technique, firm fixation has been achieved for AAD, and this technique also applies to most other patients [13, 15]. In the cases of high-riding vertebral artery or narrow vertebral pedicle, however, laminar screws are also an alternative, though they have poorer biomechanical stability than pedicle screws [16]. In this study, posterior C1-2 pedicle screw or lateral mass fixation was most often performed.

There are many classification methods for AAD, including an etiology- or dislocation direction–based method [17] that is not effective at guiding clinical treatment. In 1968, Greenberg [18] first classified AAD into two subtypes (reducible and irreducible), based on which corresponding therapeutic strategies were put forward. This classification method was a milestone, but it was too simple to fully guide the clinical approach. In 2003, based on the reduction status after skull traction and transoral anterior release, Zhu et al. [19] classified AAD into reducible dislocation, hard-to-reduce dislocation, and irreducible dislocation. This classification proved to have high practical value. As techniques have advanced, however, some cases classified as the irreducible type in the above way could be converted to the reducible type through anterior release, making Yin's classification method no longer clear [20]. In 2013, Wang [10] classified AAD into instability, reducible dislocation, irreducible dislocation, and bony dislocation and put forward corresponding therapeutic strategies: Type III cases are treated by posterior fixation and fusion after irreducible dislocations are converted to reducible dislocations by transoral release, while type IV cases can be treated by odontoidectomy. In this study, however, posterior fixation and fusion were conducted on nine cases of type III AAD following conversion to type II by trans lateral mass release (Fig. 2), four cases of type III following transoral release (Fig. 3), and two cases of type IV following transoral bony decompression and release (Fig. 4). Similar to the classification method of Wang [10], a new classification method for AAD was proposed by Tan et al. [21], which can also guide clinical practice well. However, transoral release followed by posterior fixation and fusion is also recommended for type 0 (irreducible after traction), excluding the simple posterior fixation and fusion followed by trans lateral mass release described herein. Therefore, we believe that these classification methods used for guiding therapeutic strategies can be further improved.

Effective skull traction for atlantoaxial reduction is an important way to simplify surgery and reduce complications. Except for type I cases, all cases here were subjected to postanesthesia skull traction with a maximum weight of 1/6 of body weight. The muscles, ligaments, and joint capsules blocking atlantoaxial reduction could be relaxed by traction, benefitting intraoperative reduction. The results showed that satisfactory reduction was achieved by traction in type II, and reduction exceeding 50% was achieved in nine cases of type III. Therefore, postanesthesia traction can be recommended.

The therapeutic regimen (posterior fixation and fusion) is the same for type I and type II AAD. According to one study [10], it seems that irreducible (type III) AAD originates from the unstable atlantoaxial joint. Specifically, the muscles, ligaments, and joint capsules become shorter and eventually contract due to the gradual remodeling of the lateral mass and facet joint, resulting in irreducible AAD. Therefore, attempts to adopt the strategy of posterior trans lateral mass release for type III AAD cases in this study was a right choice, which was also verified by the result that satisfactory reduction was achieved in nine cases after trans lateral mass release. Small changes in the atlanto-odontoid interspace after traction are often complicated with anterior soft-tissue contracture, so this condition should not be treated only by simple trans lateral mass release during operation, making transoral release necessary (performed in five cases in this study). Transoral plate fixation and fusion are also available for type III AAD, but this procedure often has complications such as wound infection, cerebrospinal fluid leakage, nerve injury, and internal fixation loosening [22, 23], thus restricting its popularization, and it was not done in this study. In summary, the recommended operation for type III is posterior fixation and fusion following conversion to type II by trans lateral mass release or transoral release. The possible indications for trans lateral mass release in type III AAD are as follows: (1) a high degree of reduction (> 50%) after skull traction and (2) lateral mass fusion or partial fusion seen on preoperative CT. A low degree of reduction (< 50%) after traction is considered an indication for transoral release (Fig. 7).

In this study, two cases of type IV AAD were converted to type II through transoral bony decompression and release, followed by posterior fixation and fusion. This experience suggests that odontoidectomy alone is not the only approach for type IV. Odontoidectomy may be required if reduction fails following transoral bony decompression and release.

An overview of the surgical strategies for OO with AAD is shown in Fig. 7.

In the study by Goel et al. [24], all patients showed symptomatic and clinical neurologic recovery. Other studies also obtained satisfactory efficacy. In this study, the JOA score increased from 9.58 ± 1.84 points before operation to 13.09 ± 2.68 points at 3 months after operation, 14.07 ± 2.83 points at 6 months and 14.25 ± 2.34 at 12 months after operation, all statistically significant improvements over the score before operation. The neurological recovery was obvious within 3 months after operation, and it was relatively slow at 3–6 months after operation. It can be seen that 6 months after operation seems to be the time when a plateau in recovery is reached, indicating that maximal rehabilitation within 3 months after operation is a key point. However, there were three patients whose had further neurological deterioration within 6 months after operation, which is rarely reported in the literature. One possible reason is related to the course of disease, atrophy of the spinal cord, and high T2 signals, which also support our previous findings [25] Preoperative kinematic MRI may provide guidance for these patients in determining whether there is a need to resect the posterior arch of atlas during operation [26]. Based on the above three cases, it is plausible that early intervention should also be performed on atlantoaxial instability patients without neurological dysfunction.

In this study, preoperative CT angiography (CTA) showed that the vertebral artery slightly deviated from the midline anteriorly in one case, and the head was tilted slightly to one side in the supine position during operation, causing vertebral artery injury when the lateral mass joint capsule was incised. Despite the active trans arterial embolization later, the patient died of multiple-organ failure. Therefore, it should be noted that preoperative CTA of cervical vessels is recommended, and it is necessary to carefully evaluate for abnormal vessel courses. We also recommend to pay constant attention to the presence or absence of body position changes during the operation.

Postoperative atlantoaxial infection can be catastrophic. Internal fixation can be retained in acute infection following debridement, drainage and anti-infection. In this study, one patient with infection complained of postoperative long-term post-occipital pain, which was misdiagnosed as occipital nerve neuralgia (Fig. 6). Obvious symptoms and imaging manifestations of infections were not found until 2 months after operation, at which time the internal fixation became loosened, so it was necessary to remove the internal fixation. To maintain the atlantoaxial stability, the head, neck, and cervical braces were fixed after debridement. Following infection control, nonroutine occipitocervical fusion with external fixation was performed due to a high-riding vertebral artery and destruction of the original screw track. The management of atlantoaxial infection, although rarely seen, is quite intractable, and individualized strategies are required.

Among the 56 cases enrolled, type I and type II (40 cases) were the majority, so the anterior atlanto-odontoid interspace was almost 0 mm after postoperative reduction and fixation, so it was not compared with preoperative imaging data. Similarly, the cervicomedullary angle was mostly normal before operation, so it was not included in our analyses. These are some limitations of this study. Other limitations were its retrospective nature, its small sample size, and its lack of control group.