A new classification of atlas fracture based on computed tomography: reliability, reproducibility, and preliminary clinical significance
Article information
Abstract
Study Design
A retrospective study.
Purpose
This study aimed to observe the computed tomography (CT) characteristics of atlas fracture and propose a new CT classification system, evaluate its reliability and repeatability, and discuss its clinical significance.
Overview of Literature
The treatment and classification of atlas fracture remain controversial. At present, no classification method has been standardized for atlas fracture. Injury to the transverse atlantal ligament (TAL) and lateral mass displacement is still controversial.
Methods
Seventy-five patients with atlas fracture were included from January 2015 to December 2020. Based on the anatomy of the fracture line, atlas fractures were divided into three types. Each type was divided into two subtypes according to the fracture displacement. Unweighted Cohen kappa coefficients were applied to evaluate the reliability and reproducibility.
Results
According to the new classification, 17 cases of type A1, 12 of type A2, seven of type B1, 13 of type B2, 12 of type C1, and 14 of type C2 were identified. The K-values of the interobserver and intraobserver reliability were 0.846 and 0.912, respectively, for the new classification. The K-values of interobserver reliability for types A, B, and C were 0.843, 0.799, and 0.898, respectively. The K-values of intraobserver reliability for types A, B, and C were 0.888, 0.910, and 0.935, respectively. The mean K-values of the interobserver and intraobserver reliability for subtypes were 0.687 and 0.829, respectively.
Conclusions
The new classification of atlas fractures can cover nearly all atlas fractures. This system is the first to evaluate the severity of fractures based on the C1 articular facet and fracture displacement and strengthen the anatomy ring of the atlas. It is concise, easy to remember, reliable, and reproducible.
Introduction
The atlas is located at the occipitocervical junction and is an important structure in maintaining the stability of the atlanto-occipitocervical complex. Adult atlas fractures account for 2%–13% of cervical fractures [1,2]. Despite its frequency and clinical importance, currently, no clear classification of atlas fracture has been established, and the treatment remains controversial. At present, there is no unified classification method for atlas fracture. In 1919, Jefferson was the first to introduce the classification of atlas fracture, and subsequent authors have proposed various classifications, such as the Gehweiler classification, Landells classification, Levne-Edwards classification, and Huilin Yang classification [3–8]. These classification systems are all based on the anatomy of the fracture, which cannot reflect fracture severity, and their role in guiding clinical treatment is also limited [9]. Moreover, certain atlas fractures are unclassifiable using these systems (Fig. 1). Some authors have evaluated the stability of atlas fractures according to the condition of the transverse atlantal ligament (TAL) [10,11]. However, TAL injury and lateral mass displacement are still controversial.
Demonstrated case which was not included in any existing classification. Bilateral lateral mass fracture combined with posterior arch fracture.
Given the limitations of previous classifications, this study proposed a new atlas fracture classification system based on the anatomy of the fracture line and fracture displacement. This study aimed to verify and analyze the reliability, reproducibility, and advantages of this new classification and discuss its significance for clinical treatment.
Materials and Methods
Patient data
In this study, the clinical data and computed tomography (CT) images of 75 patients with atlas fractures from January 2015 to December 2020 were analyzed retrospectively. Ethical approval for this study was obtained from the ethics committee of Ningbo No.6 Hospital (IRB no., K2020030). The patient and her family have provided consent to participate. The patient and her family have provided consent for the data in this article to be published.
The inclusion criteria were as follows: (1) patients with trauma-related atlas fracture and (2) follow-up duration of >12 months with complete radiological data. The exclusion criteria were as follows: (1) atlantoaxial malformation, (2) pathological fracture caused by infection or tumor, (3) atlantoaxial instability caused by rheumatic immune diseases, (4) incomplete radiological data, and (5) follow-up duration <12 months.
New classification system
Based on the radiological data and previous literature, this study proposed a new classification of atlas fracture [12]. According to the anatomy of the fracture line, the fracture was divided into three types. Type A is defined as the fracture line that does not involve the C1 articular facet. Type B occurs when the fracture line involves the unilateral C1 articular facet but with an intact contralateral atlas semi-ring. Type C indicates a fracture line involving the unilateral C1 articular facet but with a non-intact contralateral atlas semi-ring. These types were divided into six subtypes according to the maximum displacement of the fracture (>4 or <4 mm) (Fig. 2). Fracture displacement was defined as the minimum distance between the ends of the fracture in the maximum fracture displacement layer according to high-resolution CT scans; 4 mm is the threshold value according to the receiver operating characteristic curve analysis.
Description of the new classification of atlas fracture. A1 shows the type A1 and demonstrated case. The fracture line is not involved with the C1 articular facet, and the displacement is less than 4mm. A2 shows the type A2 and demonstrated case. The fracture line is not involved with the C1 articular facet, but the displacement is more than 4 mm. B1 shows the type B1 and demonstrated case. The fracture line is involved with the C1 articular facet, but the contralateral half ring is intact, with the displacement is than 4 mm. B2 shows the type B2 and demonstrated case. The fracture line is involved with the C1 articular facet, but the contralateral half ring is intact, with the displacement was more than 4 mm. C1 shows the type C1 and demonstrated case. The fracture line is involved with the C1 articular facet, and the contralateral half ring is not intact, with the displacement is less than 4 mm. C2 shows the type C2 and demonstrated case. The fracture line is involved with the C1 articular facet, and the contralateral half ring is not intact, with the displacement is more than 4 mm.
Radiologic evaluation
The new classification system was tested by five spine surgeons from NIngbo No.6 Hospital (three experienced and two residents in training). Each spine surgeon independently reviewed the CT images, identified whether the fracture line involved the C1 articular facet, and measured the maximum fracture displacement. To evaluate the intraobserver reliability, five spine surgeons repeated the same procedure with randomization 6 weeks after the first assessment. The results were analyzed by a different surgeon to evaluate the interobserver and intraobserver reliability. Interobserver reliability was evaluated to determine the reliability of the opinions of spine surgeons on the same occasion. Intraobserver reliability was evaluated to determine the reliability of individual surgeon’s evaluations by comparing at the different occasions for each case.
Statistical analysis
The K-values were interpreted according to the Landis and Koch criteria [13]. The result was interpreted according to the following criteria: <0, no agreement; 0.00–0.02, slight agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81–1.00, almost perfect agreement.
The paired t-test was performed to determine the significance of the differences between the mean values. The significance and power analysis were set at p-value <0.05 and 0.8, respectively. All analyses were performed using IBM SPSS Statistics for Windows ver. 26.0 (IBM Corp., Armonk, NY, USA).
Results
A total of 75 patients (male, n=44; female, n=31) were included. The mean age was 53.3±13.0 years. Of the patients, 37 (49.3%) sustained injuries following a traffic accident, 26 (34.7%) were injured by falling, and 12 (16.0%) were crushed. Twenty patients received conservative treatment, and the atlas fracture achieved good union at the last follow-up. Moreover, 55 patients underwent surgery, and 47 with less fracture displacement (≤4 mm) after surgery achieved good union. However, the remaining eight cases with greater displacement after surgery (>4 mm) showed nonunion of the atlas fracture. The mean follow-up duration was 16.4±5.1 months. According to the new classification, 17 cases of type A1, 12 of type A2, seven of type B1, 13 of type B2, 12 of type C1, and 14 of type C2 were identified.
During the two assessments, five spine surgeons performed a total of 750 classifications for the 75 atlas fracture cases. The mean K-values of the interobserver and intraobserver reliability for the new classification were 0.846 and 0.912, respectively (Table 1). The mean K-values of the interobserver and intraobserver reliability for subtypes were 0.687 and 0.829, respectively (Table 2). Thus, all K coefficients for interobserver reliability were classified as having moderate-to-substantial agreement. Most K coefficients for intraobserver reliability demonstrated almost perfect agreement, except for B1 (moderate agreement) and C1 (substantial agreement).
Interobserver and intraobserver reliability of the novel classification system
Interobserver and intraobserver reliability of subtypes of the novel classification system
Discussion
This study demonstrated almost perfect interobserver and intraobserver reliability of the new atlas fracture classification system (kappa=0.846 and 0.912, respectively). The interobserver and intraobserver reliability values for the subtypes were substantial and almost perfect (kappa=0.687 and 0.829, respectively). Our results suggested that the new atlas fracture classification system had satisfactory reliability and reproducibility. The agreement for B1 and C1 subtypes is less than that for the other subtypes. This may be because the fracture is located at the midline, which may make it difficult to ensure type B or C. Meanwhile, fracture displacement is measured manually according to the CT scan, which would have a variance in these values, perhaps resulting in greater inter- and intraobserver variations.
The new atlas fracture classification proposed herein has distinct characteristics. First, the classification can include most fracture types. Second, the classification is relatively simple and easy to remember and communicate. Third, the classification is the first to evaluate the fracture severity based on the relationship between the fracture line and atlantoaxial articular surface. C1 articular facet injury may lead to traumatic arthritis (Fig. 3) [14–16]. Therefore, this classification system can guide the prognosis in some ways.
Case example of atlanoaxial fusion after healing of atlas fracture. (A, B) showed the fracture line involved with left atlanoaxial joint. (C, D) showed the fracture after internal fixation. (E, F) showed left atlanoxial joint spontaneous fusion after 1-year follow-up (red arrow).
Li et al. [17] proposed the “buoy phenomenon” and concluded that the reduction of the displaced lateral masses to restore the C0–C2 height and maintain the ligamentous tension is the key to treating the atlas fracture. Some authors have also recommended only atlas open reduction and internal fixation (ORIF) to treat unstable atlas fractures with TAL rupture, which also achieved good clinical results [18–21]. According to the “buoy phenomenon” and experience with atlas ORIF, even in cases of TAL rupture, the union of atlas fractures without severe displacement can preserve the height of C0–C2, leading to good clinical outcomes (Fig. 4). Therefore, an intact TAL is not a decisive factor in determining the treatment approach for atlas fractures.
Demonstrative case of transverse atlantal ligament (TAL) injury. (A) showed atlas fracture with A1 subtype, combined with TAL injury (blue arrow). (B) showed atlas fracture union after 3 months with conservation treatment.
To the best of our knowledge, this study is the first to introduce a treatment algorithm for atlas fractures (Fig. 5). According to our classification system, for subtype 1, patients can be monitored for any type. A collar or brace is available for conservative treatment. If the patient cannot tolerate long-term external fixation, atlas ORIF may be performed. For subtype 2, surgery is recommended if the fracture displacement is still >4 mm after skull traction. C1 ORIF or halo-vest is an alternative option if the displacement is <4 mm after skull traction. C1–C2 fixation and fusion are recommended if the fracture displacement is still >4 mm during the operation, particularly if the patient has an anterior arch fracture. Owing to the difficulty in reducing and treating anterior arch fractures with large displacement, TAL damage may occur, and fracture nonunion results in instability (Fig. 6). Temporary C1–C2 fixation is not recommended because the rotation function may be reduced after removing the instrumentation [22]. If the fracture displacement can be reduced ≤4 mm during the operation, atlas ORIF is an alternative to preserve the C1–C2 rotation function (Fig. 7).
Treatment algorithm for atlas fractures. ORIF, open reduction and internal fixation; TAL, transverse atlantal ligament.
Case example of A2 subtype atlas fracture. (A) showed A2 atlas fracture (fracture displacement large than 4 mm; blue arrow). (B) showed fracture displacement was still larger than 4 mm after option (blue arrow). (C) showed atlas fracture nonunion after 1-year follow-up (blue arrow).
Demonstrative case of C1 open reduction and internal fixation (ORIF) after skull traction. (A) showed atlas fracture with displacement of 4.4 mm in B2 subtype. (B) showed fracture displacement was reduced to 2.92 mm after skull traction. (C) showed we performed C1 ORIF to maintain the C1–C2 rotation function. (D) showed atlas fracture union after 6 months.
This study has several limitations. First, our study only tested the reliability of the atlas fracture classification; therefore, the validity of the classification must be evaluated by comparing it with intraoperative findings and its clinical value for decision-making. Second, the study sample was small, and all study observers were in the same hospital. Future evaluations with a larger sample and experienced observers from other medical centers are warranted to confirm our classification system. Third, this study is retrospective; thus, a prospective study is needed. Fourth, this study focused only on the bone structure, indicating that the ligament complex, particularly regarding TAL stability, requires additional assessment.
Conclusions
The new classification system is the first to evaluate fracture severity by C1 articular facet and fracture displacement, to weaken the concept of TAL rupture and strengthen the anatomy ring of the atlas. This study revealed substantial inter- and intraobserver reliability. Our atlas fracture classification system is concise, reliable, and reproducible.
Key Points
This system is the first to evaluate fracture severity based on the C1 articular facet and fracture displacement and strengthen the anatomy ring of the atlas.
The new classification of atlas fracture can cover nearly all cases of atlas fractures.
This study is the first to introduce a treatment algorithm for atlas fracture.
This atlas fracture classification system is concise, reliable, and reproducible.
Notes
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Acknowledgments
This study was assisted by the Medical Records and Statistics Office of our hospital.
Funding
The Program of Natural Science Foundation of Zhejiang Province (LY18H060007), and Ningbo yingzhou district science and technology bureau social development project (2017-1-35 and 2021AS0067) funded this study.
Author Contributions
All authors participated in the management of the patient. CYL and LWJ drafted the manuscript. HXD and WY collected the clinical data. JWY and MWH supervised the case and supervised the writing of the manuscript. All authors read and approved the manuscript.