One Possible Cause for the Frequent Occurrence and Nonunion of Type II Odontoid Fracture

Background. To analyze the underlying causes of frequent occurrence and nonunion of type II odontoid fracture. Methods. CT scans along with 3D imaging software (Mimics software) were used to measure the bone density of the axis. The axis was divided into three parts, including the odontoid of the axis (the rst part), the base of the odontoid (the second part) and the body of the axis (the third part). The CT value of the axis was measured and analyzed in different axial planes from top to bottom, followed by calculation and comparison of the mean CT value of the three parts of the axis. Results. The mean CT value of the odontoid (the rst part), base of the odontoid (the second part) and body of the axis (the third part) was 651.35±188.32, 318.38±98.82 and 397.45±93.59, respectively. In addition, the interval variation of CT value of different axial planes was initially decreased and further increased with the change of axial planes from top to bottom. Conclusion. The mean CT value of the base of the odontoid was signicantly lower than that of the odontoid or the body of the axis. Therefore, the base of the odontoid was the transition region of shape and bone density, which may be one possible cause for the frequent occurrence and nonunion in the type II odontoid fracture in axis fracture. mainly and the body of the axis (651.35 ± 188.32). Govender et al [12] reported the mean difference of surface area between type-II and type-III fractures 134.72 2 , which might inuence healing after fracture. Iwanaga et al [21] and Heggeness et al [22] the odontoid process destructive forces from the apical ligament, which can lead to nonunion of type II odontoid fracture. Crockard et al [23] and Moskovich et al [24] occurrence of in type-II a of of


Background
Odontoid fracture, rst described by Mixter and Osgood in 1910, is the most prevalent fracture of the axis [1] , which is mainly categorized by Anderson and D'Alonzo classi cation [2] . In addition, the type II odontoid fracture is the most prevalent type. Anderson and D'Alonzo [2] reported 32 (54%) and Greene et al [3] revealed 120 (60%) in their studies. However, the causes for the frequent occurrence and nonunion of type II odontoid fracture are complex and unclear. To this end, the current study aimed to explore the possible cause for frequent occurrence and nonunion of type II dens fractures.

Inclusion and Exclusion Criteria
CT images of 136 patients (73 males, 63 females, age: 19-87 (mean 52.35)) were randomly collected form the CT scan pool in our institution from February 2018 to February 2019. These enrolled patients were subjected to CT scanning (0.5 mm-thick/slice) of cervical spine (Philips Brilliance 64 CT, Philips Medical Systems, Amsterdam, Netherlands). Patients were eliminated from this study if they had: (1) CT scans after operation, (2) cervical spine with other pathological diagnosis, (3) tumor or infection, and (4) abnormal cervical anatomy. Our study was approved by Medical Ethics Committee of Ningbo No.6 Hospital and all investigations were conducted in conformity with ethical principles of research.

Evaluation Of Ct Outcomes
We rst imported the volumetric imaging data preserved in digital imaging and communications in medicine format into Mimics 21.0. Software of Mimics 21.0 was used for the measurement of CT value and 3D reconstruction (Fig. 1). In terms of CT value measurement, the oval region of interest (ROI) was placed over an axial image of the axis (Fig. 1). Of note, the principle of ROI implantation was increasing trabecular bone as much as possible while avoiding the cortical bone as well as heterogeneous areas [4] . The mean CT value measured by Mimics 21.0 represented the bone density of the vertebral trabecular bone. The axis was divided into three parts, including the odontoid of the axis (the rst part), the base of the odontoid (the second part) and the body of the axis (the third part) (Fig. 2). CT values of the axis were measured and analyzed in different axial planes from top to bottom, followed by calculation and comparison of the mean CT values of the three parts of the axis. Mimics software was adopted for the calculation of CT values (accuracy: 0.01HU) and lengths (accuracy: 0.01 mm). To be speci c, CT images were magni ed three times and assessed three times by every spine surgeon. Odontoid height of midsagittal plane was de ned as H1, and the axial height of midsagittal plane was de ned as H2 (Fig. 3).
Statistical analysis SPSS version 17.0 (SPSS Inc, Chicago, IL, USA) was adopted for statistical analysis. Data were displayed as mean ± standard deviation (SD) or ranges.
Morphometric data were analyzed by Student's t test, where a P value of < 0.05 indicated statistical signi cance. And gender difference was analyzed by twosample t test with equal variances. In addition, ANOVA test along with post hoc SNK-q test was employed to compare the mean CT value of different parts for the axis.

Results
The mean CT value of the rst part (the odontoid of the axis), the second part (the base of the odontoid) and the third part (the body of the axis) of the axis was 651.35 ± 188.32 (Table 1, Fig. 4 Fig. 4), respectively. In addition, the mean CT values of the three parts of the axis was statistically signi cant in males, females and total population ( Table 2). In patients with 18-39 years of age, the mean CT value of the rst, second and third part of the axis was 725.06 ± 168.08 (Table 3, Fig. 4), 360.6 ± 82.83 (Table 3, Fig. 4) and 455.87 ± 70.34 (Table 3, Fig. 4), respectively, which was statistically signi cant in different axis parts (Table 4). In patients with 40-59 years of age, the mean CT value of the rst, second and third part of the axis was 694.23 ± 176.38 (Table 3, Fig. 4), 343.75 ± 78.64 (Table 3, Fig. 4) and 414.12 ± 81.66 (Table 3, Fig. 4), respectively, which was signi cantly different (Table 4). Similarly, in patients with 60-90 years of age, the mean CT value of the rst, second and third part of the axis was 589.89 ± 186.06 (Table 3, Fig. 4), 282.86 ± 102.71 (Table 3, Fig. 4) and 355.6 ± 90.17 (Table 3, Fig. 4), respectively, with statistical signi cance in different axis parts ( Table 4). The height of the axis at the sagittal plane was 32.44 ± 2.37 (26.98 ~ 39.09 mm) ( Table 5), and the height of the odontoid at the sagittal plane was 15.96 ± 1.22 (13.09 ~ 19.04 mm) (Table 5). Therefore, the odontoid height was approximately equal to half of the axis height. Locally weighted scatter plot smoothing (LOESS) was used to show the mean values of CT value of the axis in different axial planes. Interval changes of CT value for the different axial planes were initially decreased and further increased from top to bottom in axial planes (Fig. 5).

Discussion
Type II dens fractures account for approximately two-thirds of all odontoid fractures. Type II odontoid fractures are generally detected at odontoid-body junction base, with complicated in uencing factors. Apart from morphological appearance, the lowest bone density within the axis is likely to be another signi cant factor causing type II fractures. The current study aimed to analyze the possible causes of the frequent occurrence and nonunion of type II odontoid fracture.
The factors causing high incidence of type II odontoid fracture Odontoid fractures accounts for approximately 60% of all axis fractures and 10-18% of all cervical spine fractures [3,5,6] . The type II odontoid fracture is reported as the most prevalent dens fracture. Moreover, type II dens fractures also consists of about two-thirds of all odontoid fractures [7] . Greene et al [3] revealed 120 (60%) in their study, while Anderson and D'Alonzo [2] reported 32 (54%). Moreover, it remains controversial of the fracture etiology. Amling et al [8] reported that cortex thickness at dens base is only 35% of dens or one-third of the axis in both normal and osteoporosis individuals. According to Wang et al [9], the above anatomically distinct parameters might play critical roles in the mechanism of type II fractures. The mean CT value of the base of the odontoid (318.38 ± 98.82) was signi cantly lower than that of the odontoid (397.45 ± 93.59) and the body of the axis (651.35 ± 188.32). Moreover, odontoid base is the conjunct region between the body of the axis and odontoid, which is revealed as the transition region of shape and bone density within the axis. Amling et al [10] reported that dens base is the region of least resistance for fractures due to poorer trabecular interconnection as well as decreased trabecular bone volume. The trabecular microarchitecture further revealed that trabecular number as well as bone volume was signi cantly decreased, while trabecular separation was increased within dens base, in comparison with the dens or corpus of C2, which might explain the frequent occurrence of type II dens fractures [10] . Histological investigation revealed poor trabecular interconnection of subdental synchondrosis within surrounding cancellous bone [11] . Wang et al [9] demonstrated that the odontoid base had reduced bone volume fractions due to its micro-architectural weak points. Wang et al [9] cautiously speculated that higher probability of type II fractures at the base of the odontoid junction because of load-bearing pillars role of radial rodlike trabecular for force transmission from plate-like trabeculae running parallel to articular surfaces. In the case of convergence and transmission of energy to the weakest trabecular bone, tear would occur rst. The energy is transmitted along the torn trabecular bone to eventually cause fractures due to the heavy burden. Interval changes of CT value for the different axial planes were rst decreased and then increased from top to bottom in the axial planes. These may be the possible cause reasons for the frequent occurrence of type II odontoid fracture.
In this paper, the type II odontoid fracture was named as surgical neck fracture of the axis based on two factors. The rst reason was that type II odontoid fracture was located at the axis transition region, which looked like the neck of the axis. The second reason was the frequent occurrence of nonunion in type II odontoid fracture in the case of conservative treatment [12] . The nonunion incidence of type-II fractures following conservative therapy is reported to range between 15% and 85% [12,13] . Nonunion of type II odontoid fracture is caused by a variety of factor, among which, disrupted blood supply due to trauma has been considered to make great contribution [14,15] . Some studies have suggested that disrupted blood supply in odontoid process also contributes nonunion as well as avascular necrosis [14,16,17] . However, some anatomical researches have revealed vascular arcade of the odontoid process as well as body of the axis [18][19][20] , therefore, blood supply is not disrupted in the odontoid process in the case of type-II fractures. Some studies have shown the de cient bone mass as well as an inadequate number of trabeculae at the odontoid process base, which might be another possible reason for the frequent nonunion following type-II fractures. Amling et al [8] reported a 55% decrease on mean volume of trabeculae at the odontoid process base over the odontoid process and axis.
Amling et al [10] reported that reparative callus following fractures mainly initiated from fractured trabeculae, thus, cancellous bone recovered faster than cortical bone. Regardless of the healthy status of the bone, the bone volume is signi cantly decreased in the base of odontoid process than dens as well as the body of axis, with a lack of su cient starting point for reparative callus, which may be one of the possible reason underlying the high incidence of nonunion within the type II odontoid fracture [8] . In this study, the mean CT value of the base of the odontoid (318.38 ± 98.82) was signi cantly lower than the odontoid (397.45 ± 93.59) and the body of the axis (651.35 ± 188.32). Govender et al [12] reported that the mean difference of surface area between type-II and type-III fractures was 134.72 mm 2 , which might in uence healing after fracture. Iwanaga et al [21] and Heggeness et al [22] reported that the odontoid process was under destructive forces from the apical ligament, which can lead to nonunion of type II odontoid fracture. Crockard et al [23] and Moskovich et al [24] suggested that the occurrence of nonunion in type-II fractures as a consequence of the interposition of the transverse ligament at the site of the fracture.

Conclusion
The mean CT values of the odontoid and the body of the axis were signi cantly higher than that of odontoid base. The conjunct region between the odontoid and the body of the axis was the transition region of shape and bone density, which may be the possible cause for the frequent occurrence and nonunion in the type II odontoid fracture. Declarations Ethics approval and consent to participate Our study was approved by the Regional Ethics committee of Ningbo No.6 Hospital and all investigations were conducted in conformity with ethical principles of research.

Consent for publications
Not applicable.

Availability of data and materials
The dataset used during the current study are available from the corresponding author on reasonable request. The axis was divided into three parts. a, the rst part was the odontoid of the axis. b,the second part was the base of the odontoid. c, the third part was the body of the axis.