Abstract
-
Objective
This study aims to evaluate the clinical and radiological outcomes of posterior reduction and fusion strategies, with or without interfacet joints distraction and cage implantation, based on reducibility, in the surgical management of atlantoaxial dislocation (AAD).
-
Methods
Patients who underwent posterior reduction and fusion surgery for AAD in our institution were included. They were categorized into 2 groups based on reducibility. Japanese Orthopaedic Association (JOA), visual analogue scale (VAS), and patient-reported satisfaction scores were collected. The atlantodental interval, distance of the tip of the odontoid to Chamberlain’s line (DOCL), clivus-axial angle (CXA) and mean obliquity of the atlantoaxial articular facet (OAAF) were measured on computed tomography (CT) images. Fusion was evaluated using CT and dynamic x-rays.
-
Results
A total of 90 patients (45 males and 45 females) were included. Among them, 54 patients in the reducible group underwent direct posterior reduction and fusion, and 36 patients in the irreducible group were treated with additional interfacet joint distraction and cage implantation. All patients showed significant improvements in JOA and VAS scores postoperatively. In the irreducible group, the preoperative CXA was smaller, whereas the OAAF was greater. Receiver operating characteristic curve analysis identified optimal cutoff value of OAAF in predicting reducibility was 32.4° (sensitivity: 86.1%, specificity: 81.5%). Postoperative changes in DOCL and CXA were more pronounced in irreducible group. The fusion rates were comparable in the 2 groups (92.6% vs. 94.4%, p=0.730).
-
Conclusion
The reducibility-based posterior reduction fusion strategy achieves satisfactory clinical and radiological outcomes in the surgical management of AAD. For reducible cases, direct reduction under continuous intraoperative skull traction is preferred to minimize surgical trauma. In contrast, interfacet joints distraction and cage implantation are essential for irreducible cases. Preoperative OAAF may act as a potential predictor of reducibility.
-
Keywords: Atlantoaxial joint, Joint dislocations, Spinal fusion, Neurosurgical procedures, Internal fixators, Surgery
INTRODUCTION
Atlantoaxial dislocation (AAD) refers to pathological changes in the normal alignment and relationship between the atlas and axis due to trauma, degeneration, or congenital malformations, is a common disorder of the upper cervical spine, frequently associated with basilar invagination and other related conditions [
1,
2]. These abnormalities may cause compression of the spinal cord and brainstem leading to a range of clinical manifestations such as neck pain, neurological dysfunction, and even life-threatening complications [
3-
5]. The most effective treatment for AAD is surgical reduction and stabilization of the craniocervical junction [
6-
9]. The primary surgical approaches typically involve posterior reduction and anterior transoral release [
10,
11]. Recently lateral interfacet cage implantation through posterior approach has been reported to yield superior clinical and radiographic outcomes, particularly in cases of irreducible AAD [
12]. However, there is no consensus on the specific indications for interfacet joints distraction and cage implantation. Additionally, the preoperative assessment of the reducibility of AAD still poses challenges. Therefore, this study aims to comprehensively assess the clinical and radiological outcomes of posterior reduction and fusion strategies, with or without interfacet joint distraction and cage implantation, based on reducibility, in the surgical management of AAD through a retrospective study. At the same time, it aims to identify the potential image predictors for preoperative reducibility assessment.
MATERIALS AND METHODS
1. Patient Population
A retrospective analysis was performed on the clinical data of patients diagnosed with and surgically treated for AAD in our institution, between October 2018 and December 2023. Patients were categorized into 2 groups based on reducibility, and different surgical strategies were employed: direct posterior screw-rod reduction and fusion in reducible group; interfacet joint distraction and cage implantation followed by screw-rod reduction and fusion in irreducible group.
The study was conducted in accordance with the Declaration of Helsinki and obtained approval from the institutional review board (IRB) of Sichuan Provincial People’s Hospital (IRB No. 2025-023). Given its retrospective nature, informed consent from patients was not required.
2. Inclusion and Exclusion Criteria
The inclusion criteria were as follows: (1) The sole criteria for AAD was an atlantodental interval (ADI) greater than 3 mm among adults (≥18 years), while in younger individuals greater than 5 mm was considered abnormal. For patients with an incomplete odontoid present (os odontoideum or evidence of a prior odontoid fracture), the same criteria were adopted by measuring the distance between the inferior rim of the atlas (C1) anterior arch and the remaining attached part of the odontoid or axis (C2) anterior-superior edge [
3,
13]; (2) patients who underwent reduction and fusion surgery through a single posterior approach; (3) availability of complete preoperative and postoperative clinical outcome data for at least 12 months, including Japanese Orthopaedic Association (JOA) and visual analogue scale (VAS) scores; (4) availability of comprehensive preoperative and postoperative radiological data, including dynamic cervical x-rays, 3-dimensional computed tomography (3D-CT), and magnetic resonance imaging (MRI); (5) a minimum follow-up duration of 12 months.
The exclusion criteria were as follows: (1) presence of other severe cervical spine conditions, including cervical tumors, infections, or ankylosing spondylitis; (2) absence of critical clinical or imaging data either preoperatively or postoperatively; (3) prior history of cervical spine surgery or intraoperative conversion to a different surgical procedure; (4) patients with bony fusion or osteophytes that prevented posterior reduction and therefore underwent anterior transoral odontoidectomy or transoral release followed by posterior reduction and fusion.
3. Clinical Assessment
Clinical data, including clinical manifestations, JOA scores and VAS scores were collected preoperatively, postoperatively, and at the final follow-up. Patient-reported satisfaction was categorized into 4 levels: very satisfied, satisfied, less satisfied, and dissatisfied, as part of the final follow-up evaluation. The information was obtained through the hospital information system, telephone follow-ups, or outpatient visits.
4. Radiographic Assessment
All patients underwent preoperative dynamic lateral radiography, CT, CT angiography, 3D-CT reconstruction [
14], MRI of the cervical spine, and 3D-printed models of the craniocervical junction were created to assist surgical planning, including the precise determination of screw trajectories and thorough evaluation of interfacet device (cage) compatibility, which is essential for enhancing surgical accuracy and ensuring procedural safety.
Radiological parameters, including ADI, DOCL, C2–7 Cobb angle, CXA, axial tilt (AT) and obliquity of the atlantoaxial articular facet (OAAF) were measured on CT images. Fusion was evaluated using CT and dynamic x-rays.
Radiological parameters were measured using previously reported methods, as illustrated in
Fig. 1. Successful fusion was defined as meeting the following criteria: (1) On CT scan, there is evidence of osseous fusion between the atlas and axis [
15]. (2) In dynamic x-ray images, there is no anterior-posterior slippage or angular displacement between the C1 and C2 [
16,
17]. Image data were obtained from the picture archiving and communication system (PACS) of our institution. Radiological outcomes were independently assessed by 2 researchers (GZ and DD) using PACS software ver. 5.5.0.18051. The average values of these measurements were used in the statistical analysis. In cases of significant discrepancies, a senior surgeon was consulted to determine the final value.
The classification system was established according to 3 consecutive diagnostic tests outlined below as previously described by Wang et al. [
7].
1) Dynamic cervical radiograph
If anatomical atlantoaxial reduction could be accomplished via flexion-extension radiographs, the patients were categorized as “type I: instability.” Individuals without full reduction on dynamic radiographs were subjected to the following evaluation procedures.
2) Reconstructive CT
If osseous fusion between the C1 and C2 is identified on reconstructive CT scans, the patient shall be classified as type IV (“bony dislocation”). Only cases without radiological evidence of C1–2 osseous fusion will proceed to the subsequent evaluation step.
3) Skeletal traction test under general anesthesia
Under general anesthesia, the patient was put in prone position. Under continuous fluoroscopic guidance, graduated cranial traction was administered using a Gardner-Wells tongs, progressively increasing the applied force up to a maximum of one-fifth of the patient’s body weight. Following a 10-minute stabilization period after complete muscular flaccidity, radiographic assessment of reducibility was conducted. Those cases achieving anatomical reduction were categorized as II (reducible dislocation), while those failing to reduce were identified as type III (irreducible dislocation).
6. Grouping Criteria and Surgical Strategies
Patients were stratified into 2 surgical groups strictly based on the reducibility of the AAD, following a standardized treatment algorithm:
1) Reducible group
This group included patients with Wang type I (atlantoaxial instability) and Wang type II (reducible dislocation). These patients underwent direct posterior reduction and fusion under consistent intraoperative skull traction.
2) Irreducible group
This group consisted of patients with Wang type III (irreducible dislocation). These patients had a fixed dislocation that could not be fully reduced despite preoperative traction or the muscle-relaxant effect of general anesthesia. As a result, they underwent posterior interfacet joint release, distraction, and cage implantation at the lateral facet joints to achieve reduction.
7. Surgical Procedures
Taking a patient with atlas occipitalization as an example.
1) Posterior reduction and fusion in reducible group
For reducible group patients, the following surgical procedures were carried out. Under general anesthesia, the patient was placed in the prone position. Skull traction was implemented with a traction weight equal to 1/6 of the patient’s body weight. A straight posterior midline incision was made to expose the squamous part of the occipital bone down to the spinous process and lamina of C2. An occipital plate was placed on the squamous part of the occipital bone and fixed to the midline bone ridge of the occiput with 3 screws. Appropriately sized pedicle screws were implanted into the C2. Pre-bent titanium rods were attached to the tails of the C2 pedicle screws, and the nuts were tightened while keeping a certain distance between the other end of the rods and the occipital plate. By compressing the rods to the occipital plate and gradually tightening the nuts, the titanium rods drove the axis to rotate anteroinferiorly, thereby reducing the AAD (cantilever reduction). When intraoperative fluoroscopy showed satisfactory reduction, the nuts of the occipital plate were fully tightened. Then, bone grafting was done using local bone and allograft. A drain was placed, and the incision was closed routinely.
2) Posterior interfacet joints distraction and fusion in irreducible group
For patients in irreducible group, the previous procedures were the same as in the reducible group, including the placement of C2 pedicle screws and an occipital plate, are performed. Bilateral facet joints are exposed, and a dedicated atlantoaxial joint spreader is used to release the lateral joints. Interfacet joint cages filled with local bone are then implanted. The height of the atlantoaxial interfacet fusion cage is designated according to the preoperatively measured DOCL. The cages reduce the vertical dislocation (i.e., basilar invagination) by distracting the joints and provide immediate structural stability. Then, the same cantilever reduction procedure is performed to further reduce the horizontal dislocation. The bone grafting and closure of the incision are finished in the same manner as previously described (
Fig. 2).
Postoperative management proceeded as follows: The drainage tube was removed 48 to 72 hours after surgery, contingent upon the volume of drainage. Following drain removal, patients were mobilized. A cervical collar was prescribed for external immobilization for a duration of 1 to 2 months. Plain radiographs (anteroposterior and lateral views), 3D-CT reconstruction, and MRI of the cervical spine were obtained. The radiological imaging data and clinical results were followed up at 3 months, 6 months, 12 months, and 2 years after the operation.
9. Statistical Analysis
IBM SPSS Statistics ver. 27.0.1 (IBM Co., USA) was used for statistical analysis. For continuous variables, a paired t-test was used to compare mean values between preoperative and postoperative measurements, while an independent samples t-test was employed to compare 2 distinct groups. Repeated-measures analysis was performed on the data collected at the 3 time points within the group. Data are presented as means±standard deviation. Categorical variables are tested using the chi-square test or corrected chi-square test. Analysis of covariance (ANCOVA) was used to compare postoperative outcomes, adjusting for preoperative baseline covariates. A p-value of <0.05 was considered statistically significant.
Receiver operating characteristic (ROC) curve analysis evaluated the predictive value of preoperative OAAF for atlantoaxial reducibility, with the optimal cutoff determined by the Youden index, and sensitivity and specificity calculated.
RESULTS
1. Baseline Characteristics
A total of 113 patients underwent surgery for AAD during the study period. Among them, 15 individuals were lost to follow-up, 6 patients who lacked complete radiographic data and 2 patients whose preoperative imaging revealed bony fusion between C1 and C2 were excluded. Finally, 90 patients (45 males and 45 females) were included in the study (
Fig. 3). The mean age was 53.0±14.4 (range, 18–86) years. The average follow-up duration was 21.2±5.0 (range, 13–34) months. The primary clinical symptoms included numbness, fatigue, pain, and an unsteady gait. The JOA scores, VAS scores, age, sex, duration of symptoms, smoking history, body mass index, comorbidities such as hypertension, diabetes, and follow-up duration are summarized in
Tables 1–
3, and
Fig. 4.
Patients in reducible group show shorter surgical time (138.7±11.1 minutes vs. 190.2±19.1 minutes, p<0.001) and less estimated blood loss (106.4±50.4 mL vs. 193.8±49.8 mL, p<0.001) compared to patients in irreducible group (
Table 4).
The baseline preoperative JOA and VAS scores did not show significant differences between reducible group and irreducible group (JOA: 13.5±1.3 vs. 13.9±0.9, p=0.141; VAS: 6.6±1.9 vs. 5.9±2.0, p=0.115). Two groups demonstrated postoperative improvements in neurological function and neck pain, with no significant intergroup differences (JOA: 15.3±1.5 vs. 15.0±1.4, p=0.368; VAS:4.1±0.8 vs. 4.2±0.8, p=0.482), which were maintained at the final follow-up (JOA: 15.7±1.3 vs. 15.3±1.3, p=0.242; VAS: 2.3±1.2 vs. 2.4±1.1, p=0.765) (
Table 3,
Supplementary Table 1).
At the final follow-up assessment, the patient-reported satisfaction rates did not show statistically significant differences between the categories of very satisfied, satisfied, less satisfied, and dissatisfied (p=0.851) (
Fig. 5).
Preoperative parameters revealed that the exhibited lower CXA and greater OAAF (138.5°±17.3° vs. 145.1°±12.6°, p=0.038; 31.9°±10.2° vs. 20.0°±16.7°, p<0.001). While the other parameters including ADI, Cobb angle, DOCL, and AT were not statistically different (5.8±2.7 mm vs. 5.2±1.4 mm, p=0.21; 14.2°±14.6° vs. 13.8°±10.2°, p=0.893; 8.2±5.5 mm vs. 9.0±4.9 mm, p=0.452; 85.7°±9.7° vs. 83.4°±15.9°, p=0.405) (
Table 5).
ROC curve analysis was conducted to evaluate the predictive value of preoperative OAAF for irreducibility. The analysis yielded an area under the curve of 0.825 (p<0.001), indicating high diagnostic accuracy. The optimal cutoff value was identified as 32.4°, yielding a sensitivity of 86.1% and specificity of 81.5% (
Fig. 6).
Postoperative radiographic parameters in 2 patient groups showed significant improvement as ADI, DOCL increased, and CXA, AT decreased (p<0.05). There were no significant differences found between the 2 groups: ADI: 0.8±1.2 mm vs. 1.1±0.8 mm, p=0.22, DOCL: 4.0±4.1 mm vs. 3.1±5.1 mm, p=0.383, while CXA and AT increased to 153.2°±10.8° vs. 150.5°±15.9°, p=0.342, 93.8°±9.9° vs. 94.5°±11.5°, p=0.776, respectively. The change of these parameters maintained at the final follow-up (
Table 5).
Notably, there was no statistically significant difference in the Cobb angle between the 2 groups of patients, whether preoperatively, postoperatively, or during follow-up (p=0.893, p=0.39, p=0.07) (
Table 5).
The radiological parameters in each group showed significant improvement compared to prior measurements within each patient group. This improvement was sustained until the final follow-up (
Supplementary Table 2).
In addition, the values of changes in various radiographic parameters before and after surgery were statistically compared between the 2 groups. The decrease in DOCL and the increase in CXA was more significant in irreducible group compared to reducible group (ΔDOCL: -4.3±4.3 mm vs. -6.1±2.1 mm, p=0.017) and ΔCXA (8.0°±9.2° vs. 12.0°±8.0°, p=0.039). For the other parameters, such as ADI (-5.0±2.8 mm vs. -4.1±1.7 mm, p=0.087), Cobb angle (-0.4°±11.8° vs. 1.1°±8.9°, p=0.518), and AT (8.1°±8.2° vs. 11.0°±8.6°, p=0.109), did not show statistical significance (
Table 6). The bone fusion rate was comparable between the 2 groups of surgical patients (92.6% vs. 94.4%, p=0.730) (
Figs. 7–
11).
To address the baseline disparity in CXA, an ANCOVA was performed (
Table 7). While preoperative CXA significantly influenced the outcome (p<0.001), the adjusted analysis revealed no significant effect of the surgical group on the postoperative CXA (p=0.228, η
p2=0.017).
During the follow-up period, one patient in the reducible group experienced loss of reduction and persistent dysphagia (
Fig. 12). while another patient developed cerebrospinal fluid (CSF) leakage immediately after surgery, which was successfully managed through revision surgery, during which a dural tear was identified on the lateral side below the foramen magnum. No major complications were observed in the irreducible group.
DISCUSSION
AAD represents a surgically formidable upper cervical disorder with a diverse range of etiologies. Its pathogenesis stems from the instability of the atlantoaxial joints [
18,
19]. Over time, there is a gradual remodeling of the lateral mass facet joints, accompanied by the shortening and eventual contracture of muscles, ligaments, and joint capsule tissues, ultimately transforming reducible AAD into irreducible AAD [
1,
7]. Consequently, it is advisable to initiate surgical treatment at the early stage of AAD to avert spinal cord injury [
7,
20]. The present study introduces a surgical strategy via a single posterior approach that is strictly based on the reducibility of the dislocation.
In the current study, a total of 90 patients with AAD underwent surgical treatment via posterior reduction and fusion, either with or without facet joint distraction and cage implantation. All surgical procedures were successfully accomplished without severe complications such as vertebral artery injury or spinal cord injury, suggesting the good safety profile of this surgical approach. The VAS and JOA scores showed significant improvement in patients during the postoperative period and throughout the follow-up. Postoperative radiographic outcomes demonstrated that the vast majority of patients achieved satisfactory reduction and bony fusion.
Patients were classified into 2 groups-those with instability, reducible dislocations and those with irreducible dislocations-based on Wang’s classification through the outcomes of preoperative dynamic x-ray and intraoperative skull traction. For patients, in reducible group satisfactory clinical outcomes were achieved through direct posterior reduction and spinal fusion alone. In contrast, patients with irreducible dislocations required additional bilateral facet joints release combined with cage implantation to achieve reduction and stabilization. While the analysis of demographic characteristics (age, sex, underlying diseases) revealed no significant differences, it is crucial to recognize that the clinical baseline-specifically the reducibility of the dislocation—differed fundamentally between the 2 groups. This clinical distinction was the primary determinant for the surgical decision-making process.
However, preoperative imaging data revealed statistically significant differences in CXA and OAAF between the 2 groups, indicating that these parameters may help distinguish irreducible AAD from reducible forms, thereby influencing the choice of surgical techniques and postoperative outcomes. Specifically, patients in the irreducible group exhibited significantly higher OAAF angles compared to those in the reducible group, indicating more pronounced abnormalities in articular surface morphology among patients with irreducible AAD [
21,
22].
Our study identified a critical radiological threshold for surgical decision-making. The ROC analysis demonstrated that a preoperative OAAF exceeding 32.4° is highly predictive of irreducible dislocation (sensitivity: 86.1%, specificity: 81.5%). This suggests that an OAAF >32.4° is a robust predictor of irreducibility, warranting consideration for facet joint cage implantation.
This finding supports the hypothesis that steeper facet joint inclination, which indicates a more severe degree of AAD dislocation, suggests that the condition has reached a more advanced stage. Clinically, this threshold serves as a valuable preoperative screening tool. This quantitative criterion enhances the precision of our treatment strategy.
Yin et al. [
21], based on 3D-CT reconstruction and morphological analysis of the lateral mass joints, classified the morphology of these joints into 5 types in 2012 (type I: slight anteversion of lateral atlantoaxial articulations [LAA] without displacement of the inferior and superior facets; type II: partial displacement of bilateral facets with anteversion of LAA; type III: separation or complete displacement of bilateral facets; type IV: facets sloped dorsally). Type III or IV facet joints typically represent irreducible AAD. Chronic dislocation leads to cartilage wear, osteophyte formation, joint space narrowing, and alterations in normal anatomical structures [
1,
23]. These changes affect the sagittal plane positional relationship between the C1 and C2, increasing the OAAF while decreasing the CXA. Therefore, comprehensive preoperative evaluation, especially through 3D imaging studies, should be emphasized in clinical practice to formulate personalized surgical plans [
24], which showed a high degree of alignment with intraoperative skull traction outcome in the authors’ center (unpublished data).
Increasing the CXA to alleviate ventral spinal cord compression is one of the key objectives in the surgical reduction of AAD [
6]. It is noticed that cages placed at the lateral facet joints may impede the horizontal reduction and correction of CXA [
4] as shown in the case in
Fig. 11. However, a greater degree of CXA increase was observed in the irreducible group in the present study. We propose that this outcome may be attributed to the difference in preoperative CXA baseline between the 2 groups. Further studies involving patients with comparable baseline characteristics are necessary to more accurately assess the influence of the cage on the correction of CXA. To determine whether this was merely a reflection of the lower preoperative baseline, we performed an ANCOVA to control for heterogeneity. Results indicated that preoperative CXA was a strong predictor of postoperative CXA, and the impact of the surgical strategy was not statistically significant. In our initial cases, we discovered that cages implanted in the interfacet joints could hinder the correction of CXA and ADI. Consequently, cages with a smaller height are generally employed subsequently, which mitigates the side-effects of the cages to a certain degree.
Intraoperative skull traction plays a critical role in the treatment of AAD [
7]. It helps adjust the positional relationship between the C1 and C2, reduce the severity of dislocation, and, in cases of reducible dislocations, restore or nearly restore the normal anatomical alignment between the C1 and the odontoid process, thereby decreasing the ADI [
10]. Concurrent posterior instrumentation further restores the C1–2 complex to its physiological position by cantilever forces and stabilizes the joint structure. For irreducible dislocations, skull traction may achieve partial reduction, thus creating favorable conditions for subsequent surgical interventions. In most cases, skull traction can also modulate the atlantoaxial interspace, relieve joint locking, and facilitate facet joints release during the operation. In 2013, Wang et al. [
7] classified atlantoaxial joints into unstable, reducible, irreducible, and osseous fusion type. Intraoperative skull traction results are the key point to distinguish irreducible AAD from its reducible counterpart, providing crucial references for surgical method selection [
9,
22,
25]. In their study, anterior transoral release combined with posterior reduction and fusion was required for irreducible cases. With the advancement of interfacet joints distraction and cage implantation techniques, these cases can be successfully treated with a single posterior approach.
In the present study, the reducible group demonstrated significantly shorter operation times and reduced blood loss. This disparity is directly attributable to the additional surgical maneuvers required for the irreducible group-specifically, the bilateral facet joint exposure, comprehensive release, and cage insertion. Furthermore, the necessity for frequent fluoroscopic guidance to monitor distraction and reduction in the irreducible group inevitably extends the operative window, while the intensive manipulation of the venous plexus around the facet joints typically increases blood loss [
26].
Comparison of radiographic parameter changes from preoperative to postoperative values between the 2 groups revealed a greater decrease in DOCL and a more significant increase in the CXA in the irreducible group. This demonstrates the merit of appropriate cage implantation in reducing vertical dislocation (i.e., BI) by distracting the odontoid and C1 as a whole in a downward direction [
12,
27]. In 2004, Goel and Sharma [
28] firstly introduced posterior lateral facet release and insertion of a spacer fusion device in treatment of refractory AAD cases. Chen et al. [
12] significantly developed the technique by designing specialized instruments for facet joint release and introducing atlantoaxial facet joint cages. Furthermore, their team pioneered the concept of the posterior facet distraction and fusion technique, which has fundamentally transformed the surgical approach to irreducible AAD [
12,
27]. With the different height and lordotic angles of cages the quantitative reduction of AAD was accomplished [
29]. Studies also indicate that the use of fusion devices enhances spinal stability, minimizes the loss of reduction, prevents internal fixation failure, and improves the fusion success rate [
30]. In our center, the wide application of atlantoaxial facet joint cages has transformed most cases previously requiring a combined approach into a single posterior approach.
One case in the reducible group experienced loss of reduction, which reflects the shortcomings of simple posterior reduction and fusion, specifically the lack of anterior interfacet cage support and insufficient stability. The patient also experienced persistent dysphagia immediately after surgery. In this early case, posterior distraction between the C2 and occipital screws was conducted to achieve further reduction. This distraction may have led to increased C2–7 lordosis and the resulting dysphagia. Since then, posterior distraction has not been applied in either group of patients, and dysphagia has not been observed in any subsequent cases. Posterior distraction has been previously recognized as a primary technique for the reduction of AAD and remains in use in certain institutions. However, the high incidence of postoperative dysphagia and apnea renders this approach inadvisable [
31,
32]. Despite this isolated case of failure, while short-term clinical outcomes were generally comparable due to effective decompression in the 2 groups, the superior radiological restoration and biomechanical support in the irreducible group theoretically may safeguard against delayed myelopathy and long-term implant failure [
33]. However, it must be acknowledged that within the current follow-up period, we have not yet observed significant differences in these late-onset complications between the 2 groups. Future studies with an extended follow-up period and a larger sample size are therefore required to confirm whether these anatomical gains can be empirically translated into superior clinical benefits. A case of postoperative CSF leakage, caused by a dural tear that occurred during the suboccipital decompression procedure, was observed in the reducible group. CSF leakage is a common complication following spinal surgery, with an incidence rate ranging from 1.6% to 16.0% [
34]. Intraoperative dural injury is a direct cause of postoperative CSF leakage. Therefore, preventing dural injury during surgery and ensuring meticulous and effective repair of any damaged dura mater are critical measures for reducing the risk of CSF leakage [
35]. In most cases, when reduction was successfully achieved, suboccipital decompression is not always recommended to avoid potential complications such as CSF leakage.
The strengths of this study are as follows. First, the sample size is relatively large, which could potentially yield statistically significant results. Second, our institution is a tertiary hospital with 2 decades of experience in the surgical management of AAD through various approaches. Third, most radiographic studies were conducted within our center, making the measurement of various radiological parameters feasible.
This study also has several limitations. Firstly, we did not perform regular follow-ups 6 months after surgery, relying only on the final follow-up data for assessment. This approach failed to capture the dynamic changes that may have occurred throughout the postoperative period. Secondly, in the measurement of postoperative radiological parameters, some patients had undergone partial removal of the occipital bone as part of the decompression procedure during surgery. This made it challenging to accurately delineate the Chamberlain line and obtain reliable measurements, particularly when comparing postoperative findings with preoperative assessments. Finally, due to the retrospective nature of this single-institution study, the conclusions drawn are limited in their definitiveness. The predictive value of irreducible dislocation when the preoperative OAAF is greater than 32.4° still requires further verification. Future research should adopt large-scale, multicenter, and prospective study designs to further validate and clarify these findings.
CONCLUSION
The reducibility-based posterior reduction fusion strategy achieves satisfactory clinical and radiological outcomes in the surgical management of AAD. For reducible cases, direct reduction under continuous intraoperative skull traction is preferred to minimize surgical trauma. In contrast, interfacet joints distraction and cage implantation are essential for irreducible cases. Preoperative OAAF may act as a potential predictor of reducibility.
Supplementary Materials
NOTES
-
Conflict of Interest
The authors have nothing to disclose.
-
Funding/Support
This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
-
Author Contribution
Conceptualization: DC, QL, JL; Data curation: GZ, DD; Formal analysis: GZ, DD, JL; Methodology: GZ, HL, DD, JL; Project administration: JL; Writing – original draft: GZ, HL, DD; Writing – review & editing: DC, LC, JH, HF, QL, JL.
Fig. 1.(A and B) In the midsagittal plane of cervical computed tomography (CT), clivus-axial angle (CXA) reference to the angle between the clivus and the straight line parallel to the posterior margin of the C2 vertebral body. Cobb angle (Cobb) Cobb angle between inferior endplate of C2 and C7 vertebrae. Atlantodental interval (ADI): the horizontal distance between the anterior arch of the atlas and the dens of the axis was measured on midsaggital CT to evaluate the horizontal dislocation. The distance of the odontoid tip to Chamberlain’s line (DOCL): the shortest vertical distance from the tip of the odontoid process to Chamberlin’s line (the line connecting the posterior margin of the hard palate to the posterior border of the foramen magnum). Axial tilt (AT): The angle between the Chamberlin’s line and the posterior margin of the C2 vertebra. (C and D) The mean obliquity of atlantoaxial articular facet (OAAF) was defined as angle between the inferior articular surface of C1 and the horizontal plane in the sagittal plane.
Fig. 2.(A) Cranial traction (using Gardner-Wells tongs) was performed under general anesthesia and fluoroscopic guidance, with gradual incremental increases to 1/6 of the patient’s body weight. (B) The cages we use in our surgeries include polyetheretherketone material cages, allogeneic bone cages, and 3-dimensional-printed titanium alloy cages.
Fig. 3.The flowchart of this study. AAD, atlantoaxial dislocation; CT, computed tomography.
Fig. 4.Preoperative primary clinical symptoms of the 2 patient groups. Group A, reducible group; group B, irreducible group.
Fig. 5.Patient-reported satisfaction rate at the final follow-up. Group A, reducible group; group B, irreducible group. p=0.851.
Fig. 6.The curve illustrates the diagnostic performance of OAAF in predicting atlantoaxial irreducibility. The area under the curve (AUC) is 0.825 (p<0.001). The optimal cutoff value is 32.4°, with a sensitivity of 86.1% and a specificity of 81.5%. OAAF, obliquity of the atlantoaxial articular facet.
Fig. 7.Fusion rate at the final follow-up. Group A, reducible group; group B, irreducible group. p=0.730.
Fig. 8.A 26-year-old female patient with AAD in reducible group was successfully treated via a posterior approach without lateral cage implantation. (A–G) Preoperative images demonstrate atlas occipitalization and C2–3 fusion, findings typical of Klippel- Feil syndrome. (H–N) Postoperative imaging at 36 months following surgery shows complete reduction and solid occipitalaxial fusion. AAD, atlantoaxial dislocation; DOCL, distance of the tip of the odontoid to Chamberlain’s line; CXA, clivus-axial angle; AT, axial tilt; OAAF, obliquity of the atlantoaxial articular facet. Measurements: preoperative ADI: 4.7 mm, DOCL: 16.4 mm, CXA: 140.4°, AT: 87.3°, average OAAF: 40.6°; postoperative ADI: 0 mm, DOCL: 2.4 mm, CXA: 164.5°, AT: 98°.
Fig. 9.A 53-year-old female patient with AAD in the irreducible group was successfully treated with posterior reduction and fusion combined with lateral facet joint release and cage implantation. (A–G) Preoperative imaging revealed AAD, atlas occipitalization and C2–3 fusion, characteristic features of Klippel-Feil syndrome. (H–N) Postoperative imaging at 18-month postsurgery demonstrated complete reduction and proper placement of the interfacet cages. AAD, atlantoaxial dislocation; DOCL, distance of the tip of the odontoid to Chamberlain’s line; CXA, clivus-axial angle; AT, axial tilt; OAAF, obliquity of the atlantoaxial articular facet. Measurements: preoperative: ADI: 5 mm, DOCL: 8 mm, CXA: 137.7°, AT: 79.9°, average OAAF: 41.9°; postoperative: ADI: 0 mm, DOCL: 2.5 mm, CXA: 153.7°, AT: 94.7°.
Fig. 10.A 58-year-old female patient with AAD was successfully treated using posterior reduction and fusion combined with lateral facet joint release and cage implantation. (A–G) Preoperative imaging revealed AAD, atlas occipitalization and C2–3 fusion, which are characteristic features of Klippel-Feil syndrome. (H–N) Postoperative imaging at 22-month postsurgery demonstrated complete reduction and appropriate placement of the interfacet cages. AAD, atlantoaxial dislocation; DOCL, distance of the tip of the odontoid to Chamberlain’s line; CXA, clivus-axial angle; AT, axial tilt; OAAF, obliquity of the atlantoaxial articular facet. Measurements: preoperative: ADI: 5.2 mm, DOCL: 11 mm, CXA 126.9°, AT: 68.4°, average OAAF: 30.2°; postoperative: ADI: 0 mm, DOCL: 0.3 mm, CXA: 145.8°, AT: 90°.
Fig. 11.A 60-year-old female patient with AAD in the Irreducible group. (A–E) Preoperative imaging demonstrated AAD, atlas occipitalization, C2–3 fusion, Chiari malformation, and basilar invagination, and a markedly small CXA angle was noted. (F–J) Postoperative imaging showed effective downward movement of the odontoid process but unsatisfactory horizontal reduction after surgical intervention using the posterior facet distraction and fusion technique. AAD, atlantoaxial dislocation; CXA, clivusaxial angle.
Fig. 12.A 55-year-old female patient with AAD in the reducible group was successfully treated via a posterior approach without lateral cage implantation. (A–E) Preoperative imaging revealed AAD, atlas occipitalization, C2–3 fusion and basilar invagination. (F–J) Postoperative image showed successful reduction, and satisfactory decompression to the spinal cord. (K–O) Loss of reduction was observed at the 58-month follow-up, complete lateral joint fusion was noted. The patient experienced severe neurological deficit including walking difficulty, sensory disturbance. However, the patient refused a revision surgery, and remained under close follow-up. AAD, atlantoaxial dislocation.
Table 1.Baseline characteristics of 90 patients with atlantoaxial dislocation
Table 1.
|
Variable |
Value |
|
Age (yr) |
|
|
< 20 |
3 (3.3) |
|
20–40 |
12 (13.3) |
|
40–60 |
54 (60.0) |
|
> 60 |
21 (23.3) |
|
Sex |
|
|
Male |
45 (50.0) |
|
Female |
45 (50.0) |
|
Duration of symptoms (yr) |
|
|
< 1 |
62 (68.9) |
|
1–2 |
12 (13.3) |
|
≥ 3 |
16 (17.8) |
|
Presenting symptoms |
|
|
Numb |
51 (56.7) |
|
Fatigue |
19 (21.1) |
|
Pain |
35 (38.9) |
|
Unsteady gait |
11 (12.2) |
|
Smoking history |
28 (31.1) |
|
Body mass index (kg/m2) |
|
|
< 20 |
14 (15.6) |
|
20–25 |
37 (41.1) |
|
25–30 |
23 (25.6) |
|
> 30 |
6 (6.6) |
|
Hypertension |
11 (11.1) |
|
Diabetes |
6 (6.6) |
|
Follow-up duration (mo) |
21.2 ± 5.0 |
Table 2.Comparison of baseline characteristics between the 2 groups
Table 2.
|
Variables |
Total (n = 90) |
Reducible group (n = 54) |
Irreducible group (n = 36) |
p-value |
|
Age (yr) |
53.0 ± 14.4 |
52.5 ± 15.7 |
53.7 ± 12.3 |
0.704 |
|
Male sex |
45 (50.0) |
28 (51.9) |
17 (47.2) |
0.667 |
|
Diabetes |
6 (7.1) |
2 (4.1) |
4 (11.1) |
0.394 |
|
Hypertension |
11 (12.9) |
7 (14.3) |
4 (11.1) |
0.753 |
|
Body mass index (kg/m2) |
23.7 ± 4.6 |
23.6 ± 4.0 |
23.9 ± 5.4 |
0.748 |
|
Smoking |
28 (31.1) |
20 (37.0) |
8 (22.2) |
0.137 |
|
Medical history duration (mo) |
30.3 ± 67.3 |
38.8 ± 81.5 |
18.2 ± 37.2 |
0.162 |
|
Clinical symptoms |
|
|
|
|
|
Numb |
51 (56.7) |
29 (53.7) |
22 (61.1) |
0.487 |
|
Fatigue |
19 (21.1) |
12 (22.2) |
7 (19.4) |
0.752 |
|
Pain |
35 (38.9) |
20 (37) |
15 (41.7) |
0.659 |
|
Unsteady gait |
11 (12.2) |
8 (14.8) |
3 (8.3) |
0.515 |
|
Abnormal spinal cord signals |
67 (83.8) |
35 (79.5) |
32 (88.9) |
0.260 |
|
Occipitocervical fusion |
63 (70.0) |
36 (66.7) |
27 (75.0) |
0.398 |
|
Atlantoaxial fusion |
39 (43.3) |
25 (46.3) |
14 (38.9) |
0.487 |
|
Spinal cord cavitation |
41 (46.1) |
26 (49.1) |
15 (41.7) |
0.492 |
|
Follow-up time (mo) |
21.2 ± 5.0 |
21.9 ± 5.9 |
20.2 ± 3.2 |
0.683 |
Table 3.Comparison of clinical parameters for patients in the 2 groups
Table 3.
|
Variable |
Total (n = 90) |
Reducible group (n = 54) |
Irreducible group (n = 36) |
p-value |
|
Preoperative VAS score |
6.3 ± 2.0 |
6.6 ± 1.9 |
5.9 ± 2.0 |
0.115 |
|
Postoperative VAS score |
4.1 ± 0.8 |
4.1 ± 0.8 |
4.2 ± 0.8 |
0.482 |
|
Follow-up VAS score |
2.3 ± 1.1 |
2.3 ± 1.2 |
2.4 ± 1.1 |
0.765 |
|
Preoperative JOA score |
13.7 ± 1.2 |
13.5 ± 1.3 |
13.9 ± 0.9 |
0.141 |
|
Postoperative JOA score |
15.2 ± 1.4 |
15.3 ± 1.5 |
15.0 ± 1.4 |
0.368 |
|
Follow-up JOA score |
15.5 ± 1.3 |
15.7 ± 1.3 |
15.3 ± 1.3 |
0.242 |
Table 4.Comparison of intraoperative parameters between the 2 groups
Table 4.
|
Variable |
Total (n = 90) |
Reducible group (n = 54) |
Irreducible group (n = 36) |
p-value |
|
Operative time (min) |
159.3 ± 29.3 |
138.7 ± 11.1 |
190.2 ± 19.1 |
< 0.001 |
|
Blood losing (mL) |
141.3 ± 65.9 |
106.4 ± 50.4 |
193.8 ± 49.8 |
< 0.001 |
|
Postoperative hospital stay (day) |
5.0 ± 1.4 |
4.5 ± 1.1 |
5.9 ± 1.4 |
0.097 |
|
Fixation method |
|
|
|
0.482 |
|
Occipitocervica fixation |
54 (60.0) |
34 (63.0) |
20 (55.6) |
|
|
Atlantoaxial fixation |
36 (40.0) |
20 (37.0) |
16 (44.4) |
|
|
Bone grafting materials |
|
|
|
0.805 |
|
Autologous iliac bone |
36 (40.4) |
22 (41.5) |
14 (38.9) |
|
|
Spinous process bone+allograft bone |
53 (59.6) |
31 (58.5) |
22 (61.1) |
|
Table 5.Comparison of radiographic parameters between the 2 groups
Table 5.
|
Variable |
Preopration
|
1-Week postoperation
|
Final follow-up
|
|
Reducible group |
Irreducible group |
p-value |
Reducible group |
Irreducible group |
p-value |
Reducible group |
Irreducible group |
p-value |
|
ADI (mm) |
5.8 ± 2.7 |
5.2 ± 1.4 |
0.210 |
0.8 ± 1.2 |
1.1 ± 0.8 |
0.220 |
0.7 ± 1.1 |
1.0 ± 0.7 |
0.197 |
|
DOCL (mm) |
8.2 ± 5.5 |
9.0 ± 4.9 |
0.452 |
4.0 ± 4.1 |
3.1 ± 5.1 |
0.383 |
4.0 ± 4.1 |
3.1 ± 5.1 |
0.375 |
|
Cobb (°) |
14.2 ± 14.6 |
13.8 ± 10.2 |
0.893 |
12.4 ± 12.8 |
14.9 ± 12.1 |
0.390 |
11.1 ± 12.0 |
16.3 ± 12.8 |
0.070 |
|
CXA (°) |
145.1 ± 12.6 |
138.5 ± 17.3 |
0.038 |
153.2 ± 10.8 |
150.5 ± 15.9 |
0.342 |
150.9 ± 14.8 |
153.9 ± 10.1 |
0.301 |
|
AT (°) |
85.7 ± 9.7 |
83.4 ± 15.9 |
0.405 |
93.8 ± 9.9 |
94.5 ± 11.5 |
0.776 |
93.5 ± 10.7 |
94.8 ± 10.4 |
0.560 |
|
OAAF (°) |
20.0 ± 16.7 |
31.9 ± 10.2 |
< 0.001 |
- |
- |
- |
- |
- |
- |
Table 6.The postoperative changes of radiographic parameters compared to preoperative values between the 2 groups
Table 6.
|
Variable |
Total (n = 90) |
Reducible group (n = 54) |
Irreducible group (n = 36) |
p-value |
|
ΔADI (mm) |
–4.7 ± 2.4 |
–5.0 ± 2.8 |
–4.1 ± 1.7 |
0.087 |
|
ΔDOCL (mm) |
+5.0 ± 3.7 |
+4.3 ± 4.3 |
+6.1 ± 2.1 |
0.017 |
|
ΔCobb (°) |
–0.3 ± 10.5 |
+0.4 ± 11.8 |
–1.1 ± 8.9 |
0.518 |
|
ΔCXA (°) |
–9.6 ± 8.9 |
–8.0 ± 9.2 |
–12.0 ± 8.0 |
0.039 |
|
ΔAT (°) |
–9.3 ± 8.4 |
–8.1 ± 8.2 |
–11.0 ± 8.6 |
0.109 |
Table 7.Analysis of covariance for postoperative CXA
Table 7.
|
Source of variation |
Type III SS |
df |
Mean square |
F |
p-value |
|
Covariate |
|
|
|
|
|
|
Preoperative CXA |
9,784.23 |
1 |
9,784.23 |
160.34 |
< 0.001* |
|
Fixed factor |
|
|
|
|
|
|
Group (reducible vs. irreducible) |
90.08 |
1 |
90.08 |
1.48 |
0.228 |
|
Error |
5,308.82 |
87 |
61.02 |
|
|
|
Total |
2,097,194.67 |
90 |
|
|
|
REFERENCES
- 1. Greenberg AD. Atlanto-axial dislocations. Brain 1968;91:655-84.
- 2. Goel A. A review of a new clinical entity of ‘central atlantoaxial instability’: expanding horizons of craniovertebral junction surgery. Neurospine 2019;16:186-94.
- 3. Yang SY, Boniello AJ, Poorman CE, et al. A review of the diagnosis and treatment of atlantoaxial dislocations. Global Spine J 2014;4:197-210.
- 4. Goel A. Treatment of basilar invagination by atlantoaxial joint distraction and direct lateral mass fixation. J Neurosurg Spine 2004;1:281-6.
- 5. Liu T, Li F, Xiong W, et al. Video-assisted anterior transcervical approach for the reduction of irreducible atlantoaxial dislocation. Spine (Phila Pa 1976) 2010;35:1495-501.
- 6. Xu J, Mo S, Ma X, et al. A novel stepwise technique for safe and effective transoral release of irreducible atlantoaxial dislocation: a retrospective study of 201 cases. Spine (Phila Pa 1976) 2023;48:1148-54.
- 7. Wang S, Wang C, Yan M, et al. Novel surgical classification and treatment strategy for atlantoaxial dislocations. Spine (Phila Pa 1976) 2013;38:E1348-56.
- 8. Wang C, Yan M, Zhou HT, et al. Open reduction of irreducible atlantoaxial dislocation by transoral anterior atlantoaxial release and posterior internal fixation. Spine (Phila Pa 1976) 2006;31:E306-13.
- 9. Wang J, Xia H, Ma XY, et al. Treatment of irreducible atlantoaxial dislocation by bony deformity osteotomy, remodeling, releasing, and plate fixating through transoral approach. Int Orthop 2023;47:209-24.
- 10. Abumi K, Takada T, Shono Y, et al. Posterior occipitocervical reconstruction using cervical pedicle screws and plate-rod systems. Spine (Phila Pa 1976) 1999;24:1425-34.
- 11. Harms J, Melcher RP. Posterior C1-C2 fusion with polyaxial screw and rod fixation. Spine (Phila Pa 1976) 2001;26:2467-71.
- 12. Chen Z, Duan W, Chou D, et al. A safe and effective posterior intra-articular distraction technique to treat congenital atlantoaxial dislocation associated with basilar invagination: case series and technical nuances. Oper Neurosurg 2021;20:334-42.
- 13. Zhao D, Wang S, Passias PG, et al. Craniocervical instability in the setting of os odontoideum: assessment of cause, presentation, and surgical outcomes in a series of 279 cases. Neurosurgery 2015;76:514-21.
- 14. Su C, Chen Z, Wu H, et al. Computed tomographic angiography to analyze dangerous vertebral artery anomalies at the craniovertebral junction in patients with basilar invagination. Clin Neurol Neurosurg 2021;200:106309.
- 15. Weng C, Tian W, Li ZY, et al. Surgical management of symptomatic os odontoideum with posterior screw fixation performed using the magerl and harms techniques with intraoperative 3-dimensional fluoroscopy-based navigation. Spine (Phila Pa 1976) 2012;37:1839-46.
- 16. Yang JS, Chen H, Chu L, et al. Does additional bone grafting of atlantoaxial joint increase bone fusion rate of iliac crest autograft in posterior occipitocervical fusion? Retrospective, controlled study with 2-year follow-up. World Neurosurg 2019;125:e29-34.
- 17. Ando K, Imagama S, Ito Z, et al. Minimum 5-year follow-up results for occipitocervical fusion using the screw-rod system in craniocervical instability. Clin Spine Surg 2017;30:E628-32.
- 18. Panjabi M, Dvorak J, Duranceau J, et al. Three-dimensional movements of the upper cervical spine. Spine (Phila Pa 1976) 1988;13:726-30.
- 19. Fice JB, Cronin DS, Panzer MB. Cervical spine model to predict capsular ligament response in rear impact. Ann Biomed Eng 2011;39:2152-62.
- 20. Li C, Li L, Li Z, et al. Surgical management for posterior atlantoaxial dislocation without fracture and atlantoaxial dynamic test to confirm the integrity of the transverse ligament: a case report. Orthop Surg 2022;14:451-5.
- 21. Yin YH, Yu XG, Zhou DB, et al. Three-dimensional configuration and morphometric analysis of the lateral atlantoaxial articulation in congenital anomaly with occipitalization of the atlas. Spine (Phila Pa 1976) 2012;37:E170-3.
- 22. Liu J, Jia L, Zeng M, et al. Radiological features and internal fixation strategies of atlantoaxial dislocation combined with atlas occipitalization. Eur Spine J 2025;34:1284-94.
- 23. Goel A. Goel’s classification of atlantoaxial “facetal” dislocation. J Craniovertebr Junction Spine 2014;5:3-8.
- 24. Jian Q, Qin S, Hou Z, et al. Individualized C1-2 intra-articular three-dimensional printed porous titanium alloy cage for craniovertebral deformity. J Orthop Surg Res 2024;19:569.
- 25. Yin YH, Tong HY, Qiao GY, et al. Posterior reduction of fixed atlantoaxial dislocation and basilar invagination by atlantoaxial facet joint release and fixation: a modified technique with 174 cases. Neurosurgery 2016;78:391-400. discussion 400.
- 26. Gelinne A, Abumoussa AL, Bhowmick DA. Case report: minimally invasive modification of the Goel-Harms atlantoaxial fusion utilizing percutaneous screws and intra-articular cage is feasible and results in decreased blood loss. J Craniovertebr Junction Spine 2022;13:198-200.
- 27. Zhang B, Du Y, Zhang C, et al. Analysis of failed atlantoaxial reduction: causes of failure and strategies for revision. Orthop Surg 2024;16:2741-50.
- 28. Goel A, Sharma P. Craniovertebral realignment for basilar invagination and atlantoaxial dislocation secondary to rheumatoid arthritis. Neurol India 2004;52:338-41.
- 29. Liu Z, Zhao X, Guan J, et al. Quantitative reduction of basilar invagination: correction target of clivo-axial angle. Clin Spine Surg 2020;33:E386-90.
- 30. Zhu J, Wu J, Luo K, et al. Intraarticular bone grafting in atlantoaxial facet joints via a posterior approach: nonstructural or structural-a minimum 24-month follow-up. J Orthop Surg Res 2021;16:524.
- 31. Wang Q, Wu X, Tan M, et al. Is Anatomic reduction better than partial reduction in patients with vertical atlantoaxial dislocation? World Neurosurg 2018;114:e301-5.
- 32. Izeki M, Neo M, Ito H, et al. Reduction of atlantoaxial subluxation causes airway stenosis. Spine (Phila Pa 1976) 2013;38:E513-20.
- 33. Zheng J, Huang Z, Li K, et al. Biomechanical evaluation of anterior plate fixation with cage for basilar invagination with atlantoaxial dislocation: a cadaveric study. Neurospine 2025;22:974-86.
- 34. Barber SM, Fridley JS, Konakondla S, et al. Cerebrospinal fluid leaks after spine tumor resection: avoidance, recognition and management. Ann Transl Med 2019;7:217.
- 35. Milton R, Kalanjiyam GP, S R, et al. Dural injury following elective spine surgery - A prospective analysis of risk factors, management and complications. J Clin Orthop Trauma 2023;41:102172.
, Haotian Long1,2,*
