Traumatic posterior atlantoaxial dislocation without an associated fracture: a PRISMA-compliant case-based systematic review and meta-analysis

Article information

Asian Spine J. 2024;18(6):889-902

Publication date (electronic) : 2024 December 10

doi : https://doi.org/10.31616/asj.2024.0331

Mahmoud Fouad Ibrahim¹, Ahmed Shawky Abdelgawaad^1,2, Essam Mohammed El-Morshidy¹, Amr Hatem¹, Mohamed El-Meshtawy¹, Mohammad El-Sharkawi¹

¹Department of Orthopaedic and Trauma Surgery, Assiut University Hospitals, Assiut, Egypt

²Spine Center, Helios Klinikum Erfurt, Erfurt, Germany

Corresponding author: Mahmoud Fouad Ibrahim, Department of Orthopaedic and Trauma Surgery, Assiut University Hospitals, Assiut University Street, Assiut 71515, Egypt, Tel: +20-1142227270, Fax: +20-88-2332278, E-mail: mfibrahim@aun.edu.eg

Received 2024 August 12; Revised 2024 September 5; Accepted 2024 September 22.

Abstract

Introduction

The atlantoaxial joint has a wide range of motion, contributing to more than 50% of cervical spine rotation. Anterior atlantoaxial dislocation typically results from disruption of the transverse ligament (TL) or odontoid process (OP) deformity. Posterior atlantoaxial dislocation is primarily associated with fracture of the OP or the anterior arch of the C1 vertebra. Traumatic posterior atlantoaxial dislocation (TPAD) without an associated fracture is a rare injury. First described by Haralson and Boyd in 1969 [1], only a few cases have been documented in the literature since then [2–30]. Therefore, the pathogenesis and optimal management of these uncommon injuries remain subjects of ongoing debate.

This study commences with the presentation of a case of TPAD without an associated fracture. Thereafter, a systematic literature review and meta-analysis were conducted adhering to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Electronic databases (Scopus, PubMed, and Web of Science) were systematically searched to synthesize existing evidence on this rare injury.

Recent studies endorse the validity of case report-based meta-analyses, particularly for scenarios with limited event occurrences in traditional trial settings [31–39]. Given that comprehensive records of TPAD without an associated fracture are primarily documented in case reports, a meta-analysis based on these reports offers a justified and informative approach to synthesizing existing evidence.

This study aimed to comprehensively investigate TPAD without an associated fracture, including clinical presentation, injury mechanism, patient demographics, diagnostic methods, treatment modalities, risk factors for closed reduction failure, and clinical and radiological outcomes.

Methods

Case report

On July 2, 2020, a 21-year-old man was admitted to the trauma unit at Assiut University Hospitals following a road traffic accident involving a vehicular collision while riding his motorcycle. Following a transient loss of consciousness at the accident scene, the patient reported experiencing severe neck pain with numbness in the right upper limb upon awakening.

Upon arrival, the patient underwent a primary survey following the Advanced Trauma Life Support protocol. The cervical spine was protected using a Philadelphia collar while maintaining airway patency. His oxygen saturation was 99% in room air with no signs of respiratory distress. The blood pressure was 110/80 mm Hg and the heart rate was 90 beats per minute. His Glasgow Coma Scale score was 15/15.

The secondary survey revealed a scalp wound (sutured under local anesthesia) and severe neck tenderness with restricted movement. His neurological examination was normal (American Spinal Injury Association Impairment Scale E). Cervical spine X-ray and computed tomography (CT) revealed posterior atlantoaxial dislocation without any associated fracture. CT angiography of both vertebral arteries showed normal patency with no vascular obstruction or disruption. A cervical spine magnetic resonance imaging (MRI) revealed mild cord compression with ruptured TL (Fig. 1).

Fig. 1

(A) Lateral cervical spine radiograph illustrated posterior atlantoaxial dislocation. (B–D) Sagittal, axial, and three-dimensional (3D) reconstruction computed tomography (CT) images conclusively confirmed the diagnosis, depicting an intact odontoid process positioned in front of an intact C1 anterior arch. (E, F) 3D CT angiography revealed no evidence of vertebral artery injury or anomalies. (G, H) Magnetic resonance imaging showed a ruptured transverse ligament with mild cord compression.

The patient was transported to the operating theater for manual closed reduction, under fluoroscopic guidance, while awake. Skull traction was initiated with a 5 kg weight. The reduction was carried out by reversing the mechanism of dislocation. The patient was instructed to report any pain, numbness, or electric sensation during the maneuver. First, lateral rotation of the head was performed to place the odontoid under the narrowest part of the C1 anterior arch. Then, traction in flexion was applied, with one hand placed under the head and the other hand grasping the chin, while an assistant provided counterforce to the shoulders. When the C1 anterior arch reached the tip of the OP, as visualized on the image intensifier, the flexion of the neck allowed the OP to relocate within the C1 ring. An audible click was noticed, and the patient experienced immediate relief in neck pain. Subsequently, the angle of traction was adjusted to a slight extension (Fig. 2).

Fig. 2

Intraoperative imaging illustrates the stages of closed reduction. (A) Starting position. (B) Traction in in slight flexion. (C) The C1 arch reaches the tip of the odontoid process. (D) Flexion allowed the odontoid process to relocate into the C1 ring. (E, F) Gentle passive flexion/extension of the neck revealed marked C1–C2 instability. (G, H) Post reduction sagittal and axial computed tomography images demonstrated increased atlantodens interval.

After confirmation of reduction via image intensifier, gentle passive flexion-extension of the neck revealed marked C1–C2 instability (Fig. 2). Therefore, skull traction was continued with a reduced weight of 3 kg. A subsequent CT scan confirmed the reduction of the OP into the C1 ring, but showed an increase in the atlantodens interval (ADI) (Fig. 2). Given the instability and increased ADI, 2 days later, posterior C1–C2 fixation with fusion was performed. using C1 lateral mass and C2 pars screws with autologous iliac graft (Fig. 3).

Fig. 3

(A, B) Postoperative anteroposterior and lateral radiographs demonstrating C1–C2 fixation. (C, D) Sagittal computed tomography images show well-positioned screws and successful reduction of the odontoid process into the C1 ring, albeit with minimal joint distraction.

The patient remained neurologically intact with no complications. He was discharged 2 days postoperatively with a protective cervical collar. At the 3-month follow-up, the patient complained of bilateral suboccipital numbness, which was attributed to C2 nerve root irritation by the C1 lateral mass screws. However, his complaint was resolved spontaneously. At the 3-year follow-up, flexion-extension X-ray and CT showed stable atlantoaxial articulation with solid bony fusion (Fig. 4).

Fig. 4

At 3-years follow-up, (A, B) flexion-extension radiograph showed stable atlantoaxial articulation. (C, D) Sagittal computed tomography images revealed intact construct with solid bony fusion.

Protocol and registration

This systematic review adhered to the PRISMA guidelines and was registered in the PROSPERO online database (identifier: CRD42023308912). The review’s key components, defined using the PICOS (participants, intervention, context, outcome, study designs) acronym, are outlined in Table 1.

Table 1

PICOS acronym for the key components of this study

Eligibility criteria

We included patients with TPAD who had no associated fractures of the OP or C1 arch. Cases with dislocation due to congenital anomalies or inflammatory processes were excluded. The inclusion and exclusion criteria are presented in Table 2.

Table 2

Inclusion and exclusion criteria for this review

Quality of studies

Given the inherent bias in case reports, various tools have been used to evaluate the quality of case reports included in systematic reviews. Two reviewers independently evaluated the quality of the included reports using a standardized assessment tool [41], categorizing the studies as having low, moderate, or high risk of bias (Table 3).

Table 3

Risk of bias in studies

Results

A total of 1,625 articles were retrieved from various sources: 680 from Scopus, 505 from PubMed, 421 from Web of Science, and 19 from other sources. After excluding 477 duplicate articles, the abstracts of the remaining 1,148 articles were reviewed. Among these, 92 articles underwent full-text review, and after careful evaluation, 31 articles fulfilled the eligibility criteria and were included in the final analysis. The PRISMA flow diagram illustrates the stages of the literature selection process (Fig. 5).

Fig. 5

PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram for the stages of the selection process.

Our analysis summarizes 31 cases, including the case presented in this study (Table 4). These cases were distributed across several countries, with the majority (12 cases) in China, followed by seven in the United States, three in India, two in Iran, and one case each in Korea, Germany, Japan, Zimbabwe, Thailand, Nicaragua, and Egypt. The mean age of the patients was 41.94±13.76 years. The majority of patients were males, comprising 25 cases (81%). Traffic accidents were the leading cause, accounting for 27 cases (87%). Twenty-three cases (74%) experienced a transient loss of consciousness and 20 cases (65%) sustained facial injuries. Initially, 16 cases (52%) did not exhibit any neurological deficit. Among the 15 cases with neurologic deficit, 13 experienced complete recovery without any residual deficit, one patient experienced a fatal outcome due to an associated traumatic brain injury. For one patient, follow-up data concerning the neurological status was not available. Concerning treatment approaches, seven case (23%) were managed through closed reduction alone, 10 cases (32%) required fusion following closed reduction, and 14 cases (45%) underwent open reduction and fusion.

Table 4

Summary of cases included for review

ROC curve analysis was employed to establish the cut-off point. Using time from injury to the first attempt at closed reduction as a predictor of closed reduction failure revealed an area under the curve of 0.7870 (Fig. 6). The optimal cut-off value was 7.5 days from injury using the Youden index.

Fig. 6

Receiver operating characteristic curve for time from injury to first closed reduction attempt. The area under the curve is 0.7870.

Multivariate logistic regression analyses showed that delay in reduction (≥7.5 days) (odds ratio, 56.463; p=0.011) was a risk factor for closed reduction failure (Table 5). The Hosmer-Lemeshow goodness-of-fit test revealed no significant departure from good model fit (p=0.0743).

Table 5

Multivariate logistic regression analyses for risk factor for closed reduction failure

Discussion

TPAD without an associated fracture is a rare injury, with only 30 cases reported in the literature [1–30]. However, the actual incidence is likely to be much higher, as many cases may result in immediate mortality due to severe spinal cord distraction, thereby going unreported. Furthermore, injuries at this level are typically overlooked during routine post-mortem examinations, contributing to underreporting [1,8].

The mean age of the reported cases is relatively young (41.19±13.44 years). Older patients with degenerative cervical spine conditions likely do not survive such severe injuries [12]. Of note, 81% of the reported cases were male, possibly indicating a higher prevalence of aggressive driving among the male population, contributing to traffic accidents. Moreover, 87% of the reported cases were attributed to road traffic accidents.

The atlantoaxial joint stands out as the most mobile joint in both the spine and the entire body, responsible for more than 50% of cervical rotation [42,43]. Its stability primarily relies on the OP being securely interlocked within an osseo-ligamentous ring. This ring consists of the anterior arch of C1 anteriorly and the TL posteriorly. Consequently, anterior dislocations usually result from either fracture of the OP or disruption of the TL. On the other hand, posterior dislocation is primarily associated with fractures of the OP or the anterior arch of C1. However, posterior atlantoaxial dislocation without any associated fractures requires disruption of multiple structures, including the alar ligament, the apical ligament, the longitudinal bands of the corticate ligament, and the joint capsule. Notably, the TL may not necessarily be ruptured in such cases [8,12].

Haralson and Boyd [1] proposed extreme hyperextension with distraction as the possible mechanism of injury. This theory is supported by the high incidence (65%) of concurrent facial injuries, indicating a frontal impact. Typically, patients experience high-energy trauma from the front, resulting in extreme neck hyperextension and concurrent facial injuries [12,15,18]. Additionally, a rotational element has also been proposed to facilitate dislocation by positioning the OP anterior to the shorter paracentral part of the C1 anterior arch, reducing the critical level of distraction necessary to induce dislocation [13,20]. In cases where the TL was ruptured, we hypothesize that severe hyperflexion preceded the hyperextension injury, leading to the rupture of the TL before the occurrence of dislocation.

Patients usually present with neck pain, torticollis, dyspnea, and/or dysphagia. There was a history of loss of consciousness in 74% of cases and 68% of patients had other associated soft tissue or skeletal injuries.

Generally, dislocations are the most common injury pattern resulting in spinal cord injury, accounting for 45%–58% of cases [44]. Dislocations are inherently mechanically and neurologically unstable injuries, posing a risk of further displacement and neurological deterioration if not promptly reduced and stabilized [45]. However, TPAD without an associated fracture deviates from this trend. Remarkably, 52% of patients presented without any neurologic deficit. In patients with neurological deficits, these deficits were either transient or related to other concomitant injuries (brain injury, brachial plexus injury, or sub-axial cervical spine injury) [4,8,10–13,19,23,25,28,30].

It seems that if the spinal cord survives the initial distraction injury (which is often fatal), the patient will suffer no or mild neurological deficit. This can be attributed to two factors. First, as per the rule of Steel [46] of thirds, the spinal canal at the C1–C2 level is divided into three parts: one-third accommodates the OP, and the other two-thirds is occupied by the spinal cord and the cerebrospinal fluid (CSF). Therefore, when posterior atlantoaxial dislocation occurs, there is a substantial amount of free space before the cord is compromised [1,6,11,12,15,22,27,47]. This was supported by Tucker and Taylor [48], who illustrated that posterior atlantoaxial dislocation without an associated odontoid fracture reduced canal area by only 36%, which is sufficient to avoid cord compression. Second, in some cases of TPAD without an associated fracture, although the OP is dislocated in front of the anterior arch of C1, it remains locked in place, thereby preventing further spinal cord damage that might occur due to secondary instability [1,13,20].

A good quality X-ray can effectively diagnose TPAD. CT scans are employed to confirm the diagnosis and rule out any associated fractures. MRI can detect TL injuries; however, the associated edema and hematoma can reduce its sensitivity. MRI can also identify cord compression, hematoma, contusions, and any accompanying disk herniation. Three cases had associated vertebral artery injuries [17,28,30]. Therefore, in cases with suspected vertebro-basilar insufficiency, obtaining a CT angiography or magnetic resonance angiogram of the vertebral arteries can provide valuable information. Additionally, these imaging techniques can delineate vascular anatomy and potential anomalies, aiding in preoperative surgical planning and the choice of the safest fixation technique, taking into account the vascular context [49].

There is no clear consensus on the optimal management approach for these uncommon injuries. Closed reduction under C-arm fluoroscopy has been successful in most cases (58%). It is conducted using gradual skull traction utilizing a maximum weight of 11.8 kg, either with or without accompanying manipulation.

The technique for closed reduction comprises three distinct phases, as described by Wong et al. [6]. In the distraction phase, traction is initially employed in a mild degree of flexion to maintain the anterior arch of C1 close to the posterior surface of the OP and to prevent kinking of the spinal cord. The subsequent realignment phase involves continuing traction until the C1 ring reaches the top of the OP. The flexed angle of traction allows the C1 ring to slip forward over the OP. Then, the angle of traction is gradually shifted to a slight extension, forcing the anterior arch of C1 to come in contact with the anterior surface of the OP. Finally, the release phase entails a gradual release of traction over several hours.

It is of utmost importance to avoid over-distraction of the atlantoaxial joints and improper rotation or excessive flexion-extension maneuvers, which could result in spinal cord injury. Four patients developed neurovascular deterioration (quadriparesis, hypotension, and/or bradycardia) during traction, underscoring the procedure’s inherent risk and the need to recognize its technical complexity [8,13,18,30].

Hence, it is crucial to perform closed reduction under fluoroscopic guidance and to closely monitor the patient’s neurological status while awake. Neuromonitoring is recommended in cases where closed reduction is conducted under general anesthesia or if the patient is unconscious. This can enable the detection of any potential neurological insult early during the maneuver, allowing for prompt cessation of the procedure to reverse any deterioration.

Another critical consideration is the time elapsed from injury to closed reduction. This review identified six cases where closed reduction was delayed, either due to missed diagnosis of the injury or the patient’s unstable general condition, precluding immediate intervention. Unfortunately, delayed closed reduction failed in all six cases. Furthermore, two of these six patients developed neurological deficits during the procedure, necessitating its termination. The failure of closed reduction in these delayed cases was attributed to the formation of adhesions, scarring, and contractures [10,13,16,20,25,30].

Our analysis of these case reports confirmed that a delay in closed reduction is associated with a substantially higher risk of failure. These findings underscore the importance of timely reduction of these injuries, as the risk of failure and neurological deterioration becomes significantly elevated if closed reduction is attempted late. However, we acknowledge that individual patient circumstances and clinical judgment are paramount in determining the appropriate timing for intervention. The determination of specific time frames, such as the observed trend around 7.5 days, should be cautiously interpreted in the context of overall patient management.

Following closed reduction, verifying the stability of the atlantoaxial articulation is crucial. In our review, ten cases exhibited residual instability after successful closed reduction, necessitating additional fixation and fusion. Sun et al. [25] reported a case of re-dislocation 28 days after initially successful closed reduction, necessitating open reduction. In seven cases, the atlantoaxial articulation remained stable after closed reduction, obviating the need for supplementary surgical fixation or fusion [2,3,7,9,12,14,23]. This stability was ascribed to the intactness of the TL, which imparts adequate stability to the atlantoaxial articulation after the reduction of the OP into the osseo-ligamentous ring. The integrity of the TL was confirmed using MRI in the case reported by Chaudhary et al. [12]. However, assessing TL integrity via MRI can be challenging in traumatic settings due to concurrent edema and hematoma. To evaluate residual instability, some authors recommend controlled gentle flexion-extension of the cervical spine under fluoroscopic guidance [6,9,26,27]. However, this maneuver carries significant risks and demands extreme caution. Sun et al. [25] suggested that an ADI of ≥5 mm on standard X-rays after closed reduction indicates residual instability, necessitating further surgical fixation.

Open reduction is recommended in the following circumstances: (1) when closed reduction fails to reduce the dislocation; (2) if closed reduction is aborted due to neurovascular deterioration during the maneuver; (3) when closed reduction cannot be attempted due to significant preexisting neurological deficits, signifying a loss of neurological function reserve; (4) in instances where the patient is unconscious and neuromonitoring resources are inaccessible; and (5) old injuries, in which substantial swelling/scarring could occur, thereby rendering closed reduction challenging.

Open reduction was carried out in 14 instances, accounting for approximately 45% of the cases. Among these, partial odontoidectomy with or without excision of the C1 arch was done through the transoral approach in seven cases [4,8,10,13,18,22,30] and through the anterior retropharyngeal approach in four cases [15,16,19,20]. In the remaining three cases [11,25,29], open reduction was achieved through the posterior approach by gentle caudal traction on the spinous process of C2.

One of the seven patients who underwent open reduction through the transoral approach developed CSF leak and meningitis, which was managed by antibiotic therapy with paraventriculostomy and lumbar drainage [30]. Two of the four cases that underwent open reduction through the anterior retropharyngeal approach developed postoperative neuropraxia of the hypoglossal and superior laryngeal nerves. This was likely attributable to soft tissue retraction but resolved completely within 6 weeks [15,20].

Fixation and fusion were necessary to manage residual atlantoaxial instability in 24 cases. Various techniques have been employed for this purpose, including the posterior wiring “Gallie technique,” posterior C1–C2 transarticular screws “Magerl technique,” anterior C1–C2 transarticular screws, posterior C1–C2 screws “Goel-Harms technique,” and occipito-cervical fusion. According to multiple studies, posterior C1–C2 screw fixation is the preferred fusion technique due to its exceptional safety profile and biomechanical stability. This approach yields high fusion rates without requiring rigid postoperative immobilization and preserves motion at the atlanto-occipital joint, offering an advantage over occipito-cervical fusion [47,50–53]. Nevertheless, occipito-cervical fusion is indicated in the presence of concurrent congenital atlanto-occipital assimilation, deficient posterior arch of C1, or atlanto-occipital instability. Atlanto-occipital instability should be suspected and investigated in cases of TPAD without an associated fracture, given the close anatomical proximity of these regions. Two cases of TPAD without associated fractures reported by John et al. [28] and Peterson et al. [24] exhibited concurrent atlanto-occipital instability. Such cases often present with neurological deficits or respiratory distress due to instability at the atlanto-occipital junction. This instability was attributed to the rupture of the occipitoatlantal capsular ligaments, the alar ligaments, the apical ligament, and the cruciate ligament [54]. Consequently, these cases necessitated occipito-cervical fusion to restore stability.

Clinical outcomes revealed remarkable recovery rates among patients with neurological deficits. Of the 15 cases presenting with neurological deficits, 12 patients (80%) achieved full recovery with no residual deficits. Notably, 10 of the 15 cases with neurological deficits were attributed to concomitant injuries (traumatic brain injuries, vertebral artery compromise, concomitant sub-axial cervical spine injuries, and brachial plexus injuries). This underscores the significance of early recognition and management of concomitant injuries to ensure successful treatment outcomes and optimal recovery from neurological deficits.

Radiological outcomes revealed successful fusion in 17 of the 24 patients who underwent fixation, demonstrated by the absence of movement at the C1–C2 joint on flexion/extension radiographs. However, fusion status data were unavailable for the remaining seven patients who underwent fixation.

Among the seven patients managed conservatively without fixation, five did not exhibit any residual C1–C2 instability after closed reduction. This outcome suggests that conservative management can be effective in select cases, particularly when closed reduction achieves stable atlanto-axial alignment. However, data regarding the stability of the atlantoaxial joint in the remaining two patients were unavailable.

The main limitation of this study is the lack of higher-level evidence, as the majority of documented cases are presented as case reports without control groups. Consequently, this meta-analysis is based exclusively on case reports, which may introduce bias and compromise the accuracy of scientific findings due to the propensity to showcase more favorable outcomes in such reports. Additionally, our study employed ROC curve analysis and multivariate logistic regression to evaluate the time from injury to the first attempt at closed reduction as a predictor of closed reduction failure. However, the intensity of effort exerted during closed reductions and associated skull trauma, which could impact the feasibility of closed reduction, were not explicitly accounted for in the analysis. These factors may limit the applicability of our findings. Despite these limitations, we believe this systematic review and meta-analysis would increase awareness among practicing surgeons of the importance of early and proper management of this rare injury and could positively impact the outcome of these patients.

Based on this study, we propose the following algorithm for the management of TPAD without an associated fracture (Fig. 7).

Fig. 7

Algorithm for management of traumatic posterior atlantoaxial dislocation without an associated fracture.

Conclusions

TPAD without an associated fracture is a rare and challenging injury in the realm of spinal trauma. Patients typically present without neurological deficits or with transient or concomitant injury-related deficits. Once the patient is hemodynamically stable, closed reduction under fluoroscopy should be promptly performed while monitoring neurological status. C1–C2 fusion is reserved for cases with MRI-confirmed TL rupture or residual instability. If closed reduction fails to achieve anatomical realignment, open reduction with C1–C2 fusion is recommended. The preferred fusion technique involves posterior C1–C2 screws. This approach ensures a high level of safety and biomechanical stability.

Article information Continued