Introduction
Subaxial cervical facet dislocation with interlocked facets are high-energy injuries that can cause severe instability and often result in neurological compromise [1]. Rapid alignment restoration is essential to prevent secondary cord injury and enable definitive stabilization. Closed cranial traction is widely used because it is noninvasive and can achieve timely reduction; however, success rates vary considerably across studies, largely reflecting differences in injury morphology and reduction protocols [2,3].
The Allen–Ferguson classification describes cervical spine injuries based on the underlying mechanism (flexion, extension, compression, or distraction). Among these, compression–extension (CE) injuries represent an extension–distraction pattern characterized by posterior ligamentous disruption, facet impaction, vertical locking, and interposed capsular or bony structures. These features create a mechanically locked configuration that may resist traction-based reduction [4–7]. Prior studies have suggested an association between morphological locking and reduction failure [7], but whether CE morphology independently predicts irreducibility has not been systematically evaluated within a consecutive clinical cohort.
Other factors, including bilateral dislocation, severe translation, facet fracture, or delayed reduction, have been proposed as predictors of irreducibility, although existing evidence is inconsistent and is often limited by heterogeneous morphological definitions [8,9]. A simplified approach that dichotomizes injuries into a CE versus non-CE pattern may improve reproducibility and provide a more practical framework for identifying patients at risk of a failed closed reduction in an emergency setting.
Vertebral artery injury (VAI) is frequently associated with cervical dislocation [10]. Sudden alignment correction during traction may theoretically worsen VAI by stretching or kinking the artery [10,11]. However, the relationship between VAI and reduction success or safety has not been clarified, and, to the best of our knowledge, no previous study has directly evaluated its impact [11].
This study aimed to determine which morphological factors independently predict failed closed reduction in subaxial cervical facet dislocations with interlocked facets, with a particular focus on CE morphology, facet fracture, and bilateral dislocation. A secondary objective was to examine whether VAI influences the safety or efficacy of closed reduction. We hypothesized that a simplified CE versus non-CE classification would provide predictive value while improving clinical applicability compared with detailed mechanistic subtypes.
Materials and Methods
Study design and setting
This retrospective, cohort study was conducted at a single tertiary trauma center and was approved by the Institutional Review Board (IRB) of Saitama Medical Center. (IRB approval number: SO 2025-060). Written informed consent was waived because only de-identified clinical data were analyzed.
Inclusion and exclusion criteria
Inclusion criteria were as follows: (1) Radiologically confirmed subaxial (C3–C7) cervical dislocation with interlocking facets was present. (2) An attempt at closed reduction using cranial traction was performed.
Exclusion criteria were as follows: (1) death before any attempted reduction; (2) patients where closed reduction was not attempted due to marked instability or associated injuries requiring immediate open surgery (including skull fractures precluding halo application and hemodynamically unstable patients with severe polytrauma requiring priority lifesaving interventions; and (3) incomplete radiological or clinical documentation.
Screening process
Between January 2017 and June 2025, 123 cervical dislocation events were screened with a total of 12 events being excluded. Five patients died before reduction, four had marked instability prompting immediate open surgery, two had skull fractures contraindicating cranial traction, and one had missing radiological or clinical data.
The final cohort consisted of 111 dislocation events in 110 patients. One patient sustained dislocations at two different levels during the same traumatic event, and these were included as independent events. The overall screening and inclusion process is summarized in Fig. 1.
Outcome definition
A failed closed reduction was defined as the inability to achieve complete facet realignment under traction, thereby necessitating open reduction. Partial or unilateral realignment was also classified as failure to ensure consistent outcome categorization. Successful reduction was defined as a complete and stable realignment achieved by traction alone, as confirmed on radiographs or fluoroscopy, and maintained intraoperatively.
Variables collected
Variables collected included age, sex, dislocation level, unilateral versus bilateral dislocation, facet fracture, injury morphology, American Spinal Injury Association Impairment Scale (ASIA) grade, presence of VAI, management of VAI including embolization, and the time from injury to reduction (continuous and categorized).
Imaging and injury morphology
All patients underwent CT and CT angiography (CTA) on admission. VAI was defined on the CTA as luminal narrowing, occlusion, or pseudoaneurysm based on prior criteria [10,11]. When unilateral Grade IV VAI was detected, coil embolization was performed prior to reduction to prevent a distal embolism.
Morphology was classified according to the Allen–Ferguson system and dichotomized as CE or non-CE for analysis. This binary scheme was adopted a priori because a detailed subtype classification has only moderate interobserver reliability on CT [12]. Morphology was independently evaluated by the first author and two spine surgeons, with disagreements resolved by consensus review.
Reduction protocol
Closed reduction followed a standardized institutional protocol informed by prior reports [3].
Awake traction (preferred approach)
After applying a halo crown (ReSolve Halo System; Össur, Reykjavík, Iceland), traction was initiated at 5 kg and increased in 2–5 kg increments. Portable lateral radiographs were obtained at each increment. Based on established safety limits, the maximum traction weight was limited to 30 kg, consistent with our institutional protocol [13].
General anesthesia traction (when awake traction was contraindicated)
If awake traction was unsafe due to polytrauma, agitation, or respiratory compromise, traction was performed under general anesthesia using Mayfield fixation. Neurological monitoring consisted of continuous clinical assessment by spine surgeons and emergency physicians, and included observation of spontaneous limb movement, abnormal posturing, pupillary changes, and autonomic instability. Neurophysiological monitoring via somatosensory evoked potential and motor evoked potential was not used as it was not part of the institutional protocol.
Criteria for terminating traction and converting to open reduction
Traction was abandoned under any of the following conditions: (1) no radiographic improvement despite incremental increases in weight of traction, (2) worsening pain or neurological symptoms during awake traction, (3) cardiorespiratory instability, and (4) reaching 30 kg of traction without realignment.
Surgical management
All patients proceeded to definitive surgical stabilization after a failed attempted closed reduction. At our institution, posterior fixation is the preferred approach for the following reasons: (1) preservation of facet joint function after implant removal, (2) frequent coexistence of multilevel cervical stenosis in the Japanese population which favors posterior decompression [14], and (3) clinical guidelines recommending multilevel posterior decompression for severe spinal cord injury (ASIA class A or B) [15].
Anterior cervical discectomy and fusion were performed selectively as a staged procedure during the same hospitalization when preoperative magnetic resonance imaging (MRI) demonstrated significant disc herniation, or when anterior reconstruction provided a biomechanical advantage [16]. When closed reduction failed, posterior open reduction and fixation were performed using standard techniques, including partial facetectomy or laminar distraction, as needed.
Statistical analysis
Analyses were performed using R ver. 4.3.1 (R Foundation for Statistical Computing, Vienna, Austria). Two-sided p-values <0.05 were considered statistically significant. Continuous variables were summarized as mean±standard deviation or median interquartile range (IQR). Categorical variables were compared using chi-square or Fisher’s exact tests, and continuous variables were compared using the Mann-Whitney U test.
Time-to-reduction was analyzed categorically (≤6 hours, 6–12 hours, 12–24 hours, >24 hours) because the distribution was highly right-skewed and influenced by delayed diagnosis or referral from distant facilities. Exploratory analyses treating time as a continuous or log-transformed variable produced unstable estimates; therefore, categorical thresholds were used.
Variables with p<0.10 on univariate analysis were included in the multivariate logistic regression model and were restricted to three covariates to avoid overfitting. Multicollinearity was assessed using variance inflation factor (<2.0) and Cramér’s V (<0.7). Sensitivity analyses using alternative covariate sets were performed to confirm robustness. Model performance was evaluated using area under the curve (AUC) with bootstrap 95% confidence intervals (CIs) (1,000 iterations), and calibration was assessed using the Hosmer–Lemeshow test.
Results
Baseline characteristics
A total of 111 dislocation events in 110 patients were included. One patient sustained two-level dislocations during a single traumatic event, which were analyzed as independent events. The median age was 65 years (IQR, 52–75), and 70.3% of events occurred in men. Dislocations most commonly involved C5 (37.8%) and C6 (31.5%). Bilateral facet dislocation occurred in 56.8%, facet fractures in 47.7%, and CE-type morphology in 20.7%. ASIA A/B injuries accounted for 44.1% of events. VAI was detected in 26.1%. Time-to-reduction was ≤6 hours in 43.2%, 6–12 hours in 27.9%, 12–24 hours in 7.2%, and >24 hours in 21.6% (Table 1).
Reduction outcomes
Closed traction successfully reduced 77 of 111 events (69.4%), and 34 events (30.6%) required open reduction. Among the 63 bilateral dislocations, 38 (60.3%) were successfully reduced and 25 (39.7%) failed. Of the 25 failures, 20 displayed no unlocking, five achieved only partial or unilateral reduction, and all required open reduction. No patient exhibited neurological deterioration during traction. No newly developed VAI was suspected clinically after traction in any patient.
Predictors of failed closed reduction
On univariate analysis, failed reduction was associated with younger age (p=0.047), CE morphology (p<0.001), facet fracture (p<0.001), bilateral dislocation (p=0.020), and C7 dislocation (p<0.001). Multivariate logistic regression identified: (1) CE morphology (OR, 31.7; 95% CI, 7.1–140.9; p<0.001) and (2) facet fracture (OR, 7.8; 95% CI, 2.3–27.0; p=0.001).
Age was no longer significant on multivariate analysis. CE4 and CE5 had failure rates of 84.2% and 100%, respectively, supporting the strong association between CE morphology and irreducibility. Final predictors are shown in Fig. 2, and complete statistical results are presented in Table 2. Sensitivity analyses using alternative covariate sets, including models incorporating the C7 level, yielded consistent results. CE morphology and facet fracture remained significant predictors, while C7 dislocation showed only a non-significant trend (Table 3).
Vertebral artery injury
VAI was more common in the successful reduction group than in the failure group (31.2% vs. 14.7%, p=0.10) and did not independently predict the reduction outcome in any model.
Timing of reduction
The median time to performing the reduction was five hours in the success group and 6 hours in the failure group. Categorized time-to-reduction was not associated with outcome (p=0.30), with success rates of 70.8%, 61.3%, 100%, and 66.7% across the four-time categories. Delays primarily reflected a delay in diagnosis or a prolonged transfer time from an outside hospital.
Dislocation level
C7 dislocation was strongly associated with a failed reduction (p<0.001), while a C5 dislocation was more frequent among successes (p=0.002). Because the level of dislocation correlated with CE morphology and facet fracture, it was excluded from the multivariate model.
Neurological status (ASIA grade)
ASIA A/B injuries were more common in the failure group (52.9% vs. 40.3%) but they were not statistically significant (p=0.22).
Model performance
The multivariate model demonstrated good discrimination (AUC, 0.86; 95% CI, 0.77–0.94) and acceptable calibration (Hosmer–Lemeshow; χ2=3.7, p=0.89).
Representative case 1. Successful closed reduction in non-CE injury
A 30-year-old man sustained a C5/6 dislocation during a backward somersault. He presented 1 hour after injury with complete tetraplegia (ASIA A) (Fig. 3A). CT angiography showed no VAI. Closed traction was initiated 1.5 hours after injury and achieved complete facet realignment within 2 hours (Fig. 3B). Due to complete paralysis, MRI was omitted, and C3–7 laminoplasty with C5/6 posterior fusion was performed the same day (Fig. 3C).
Representative case 2. Failed closed reduction in CE5 injury
A 69-year-old man fell from the 14th floor and arrived in hemorrhagic shock with a Glasgow Coma Scale score of 3. Initial ASIA grading was not feasible due to depressed consciousness. Following early resuscitation, his level of consciousness improved sufficiently to allow for a neurological assessment, which revealed ASIA A with a sensory–motor level at C4. CT demonstrated burst fractures of C1–C3, severe anterior translation of C7 with bilateral facet fractures consistent with CE5 morphology, and an associated T1 vertebral body fracture (Fig. 4A–C). Halo traction was increased in a stepwise fashion to 30 kg; however, no facet unlocking occurred, and only minimal sagittal correction at the vertebral body level was achieved (Fig. 4D).
Due to persistent interlocking, a posterior open reduction with decompression and posterior fixation from C5 to T2 was performed (Fig. 4E). Subsequent stabilization of the upper cervical fractures (C1–C3) was performed separately and was unrelated to the reduction maneuver.
Discussion
This study demonstrated that CE-type injury and facet fracture were strong, independent predictors of failed closed reduction in subaxial dislocations with interlocked facets, whereas bilateral dislocation, VAI, age, and time-to-reduction were not. These results reinforce the concept that morphology-specific structural disruption, rather than generic markers of instability, principally determines reducibility [7].
CE injuries produced by extension–distraction forces disrupt the posterior ligamentous complex and create vertical dissociation and facet impaction. These features inherently resist traction-based unlocking [4–6]. This mechanism aligns with the markedly elevated failure rates in CE4 and CE5 injuries and supports prior reports cautioning against prolonged traction when morphological locking is evident [7,17].
Facet fracture was the second independent predictor of irreducibility. Fractures of the articular processes can wedge between or vertically impact the facets, thereby mechanically preventing realignment [6]. Clinical studies have associated facet fractures with a reduced success of closed reduction and a higher need for posterior or combined approaches [1]. Accordingly, the identification of CE morphology or facet fracture on admission CT should prompt early consideration of open reduction.
Bilateral dislocation was not independently associated with failure after adjustment. This suggests that bilaterality functions largely as a surrogate for CE-type disruption or facet fracture rather than as an intrinsic predictor of traction resistance [7,16].
VAI did not predict reducibility, and its higher prevalence in reducible flexion-type injuries likely reflects the injury mechanism rather than any effect on traction outcomes [10,11]. Thus, VAI should be considered primarily as a vascular risk factor rather than a determinant of mechanical reducibility.
A simplified CE is a non-CE morphological classification that was adopted a priori because detailed Allen–Ferguson subtypes show only moderate interobserver reliability, particularly on emergent CT imaging [12]. The binary classification preserved prognostic strength while improving reproducibility and applicability for rapid decision-making in an emergency setting.
Younger age and C7 involvement were associated with failure on univariate testing but lost significance after adjusting for morphology. This indicates that these variables primarily act as markers of CE injury patterns and facet fractures rather than independent predictors. This is consistent with the biomechanical vulnerability of the cervicothoracic junction to extension–distraction failures [4–6].
Although delayed reduction has traditionally been considered a risk factor for irreducibility, the present study did not identify a significant association between time-to-reduction and reduction outcomes. This likely reflects the heterogeneity of clinical pathways in a tertiary referral setting, where delays often arise from late diagnosis or interhospital transfer rather than from prolonged traction attempts. Additionally, most reductions in this cohort were initiated within the early postinjury period, potentially limiting the ability to detect time-dependent effects. Therefore, timing should not be interpreted as a primary determinant of reducibility once morphological factors, such as CE injury and facet fracture, are considered.
Clinically, these findings support limiting prolonged traction attempts when CE injury or facet fracture is evident, and early transitioning to posterior or combined open reduction should be considered. When traction is attempted, adherence to established safety practices, including incremental loading, awake monitoring, and stepwise imaging, remains essential [3].
A practical implication of these findings is an algorithmic approach in which CE morphology or facet fracture prompts limiting traction to restoring sagittal alignment only, without insisting on facet unlocking, and transitioning swiftly to open reduction.
Conclusions
Closed cranial traction reduced approximately two-thirds of subaxial cervical dislocations with interlocked facets. CE-type morphology and facet fracture were strong, independent predictors of a failed reduction, while bilateral dislocation, VAI, and reduction timing were not. Early recognition of CE morphology or facet fracture should prompt timely transition to open reduction, and the simplified CE versus non-CE framework offers a practical tool for rapid decision-making in an emergency setting.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
