Introduction

Odontoid fractures are the most common cervical spine fractures in adults aged 65 years or above and show an increasing incidence with age [1, 2]. These fractures constitute 9–15% of all adult cervical spine fractures [3]. In recent years, several large-scale epidemiologic studies from countries such as the US and Norway have characterized the demographic patterns associated with odontoid fractures [2, 4, 5]. Despite this, nationwide studies containing data from fracture-specific registers are currently lacking. In light of this, the increasing incidence of odontoid fractures paralleling with an aging population poses a substantial economic and social burden on health care systems globally. Consequently, updated epidemiological investigations are necessary to improve the understanding of odontoid fracture epidemiology and management. The aim of this study was to perform an epidemiological survey of patients with odontoid fractures in terms of demographic, injury-related and surgical characteristics using the Swedish Fracture Register (SFR).

Methods

This study was reported according to STROBE guidelines for observational studies.

The Swedish Fracture Register (SFR)

This study is based on the SFR, a nationwide quality register collecting data pertinent to all fracture types of the extremities, pelvis and spine in Sweden. The SFR was established in 2011 and comprises data from over 900,000 patients [6]. Spinal fractures have been registered since 2015. From the inception of the SFR, coverage has grown to include all the 54 orthopeadic departments in Sweden, demonstrating nationwide coverage in 2020 [7]. Inclusion in the register is limited to fractures sustained within Sweden in patients with valid Swedish social security numbers and a fracture diagnosis confirmed with radiographic imaging. The physician records the date and mechanism of injury and classifies the fracture based on clinical and radiographic characteristics. Fractures in the SFR may be registered by the evaluating physician in the Orthopeadic emergency department or by the treating surgeon if the fracture is treated surgically.

Study population and eligibility criteria

Prospective register data was collected from the SFR between February 2015 and September 2021. Patients ≥ 18 years with a fracture of the odontoid process (Anderson and D’Alonzo classification [8]; Fig. 1) were eligible for inclusion. Demographics including age and sex as well as injury and treatment related data encompassing injury date, injury mechanism, injury type, fracture type (type I, II and III), neurological deficits, and treatment type were queried. Additionally, the diagnosis of osteoporosis (M80, M81, M82) was retrieved from the Swedish National Patient Register.

Fig. 1
figure 1

Reprint of this figure was permitted by the Swedish Fracture Register. Please note that ‘’traumatisk spondylolistes’’ corresponding to traumatic spondylolisthesis is not included in the analysis of the current manuscript

Classification of C2 fractures in the Swedish Fracture Register according to the web-based protocol at https://sfr.registercentrum.se/.

Full size image

Statistical analysis

Statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). Continuous data are presented as mean (standard deviation [SD]) and median (min-max)). For categorical variables, count (n) and proportion (%) are reported. Differences between groups were assessed using one-way analysis of variance (ANOVA) for continuous variables and using chi-square test for non-ordered categorical variables. For pairwise comparisons between groups, the Fisher’s exact test (2-sided) was calculated for dichotomous variables and ANOVA for continuous variables. A two-tailed p-value < 0.05 was determined statistically significant.

Results

Demographic characteristics

Totally, 1,154 odontoid fractures were identified in the SFR (Table 1). The cohort comprised 30 patients with type I, 583 patients with type II and 541 patients with a type III fracture. Figure 2 displays the seasonal variation in injury incidence. The mean [SD] age was 77.2 (13.8) years, with type II fractures being oldest at 78.2 (13.1) years (p < 0.001, Fig. 3). Sex distribution was similar in the total study group, but significant differences were found between type II and type III fractures as more males suffered a type II fracture compared to type III fractures (56% vs. 45% p < 0.001, Table 1).

Table 1 Demographic characteristics of 1,154 patients with odontoid fractures
Full size table
Fig. 2
figure 2

Seasonal variations of odontoid fractures. Vertical lines denote the confidence intervals

Full size image
Fig. 3
figure 3

Age at injury by sex

Full size image

The presence of neurological deficits as well as a diagnosis of osteoporosis did not differ between the three fracture types. Moreover, most patients underwent non-surgical treatment (81%). Surgical treatment was more prevalent with type II fractures (22%) compared with type III fractures (17%, p = 0.02).

Degree of trauma and injury mechanism

Only 135 (14%) patients were involved in a high-energy trauma, while 800 patients (86%) had a low impact trauma in the injury mechanism (Table 2). Significantly more patients in the type II fracture group (89%) were recorded to have a low-energy trauma compared with type III fractures (82%, p < 0.01). There was no difference between the prevalence of the three fracture types when all injury mechanisms were combined (fall from standing height or less, fall from height, motor vehicle accident, unspecified fall, or other causes). However, significantly more patients in the type II fracture group (66%) sustained a fall from standing height or less compared to patients with type III fractures (58%, p = 0.01). Conversely, a greater proportion of patients with type III fractures (12%) were involved in motor vehicle accidents than type II fractures (8.1%, p = 0.04).

Table 2 Type of trauma and injury mechanism
Full size table

Injury characteristics by age

Most of the patients aged 18–30 years (75%) were involved in a high-energy trauma in comparison to the age groups 31–65 years (38%, p < 0.01) and > 65 years (9.6%, p < 0.001, Fig. 4, Supplementary Tables 1a-c). Additionally, significantly more patients aged > 65 years (67%) had fallen from standing height or less compared to those aged 18–30 years (23%, p < 0.001) and 31–65 years (37%, p < 0.001). Patients aged 31–65 years (20%) were more likely to fall from a height compared with the > 65 years group (13%, p = 0.04). In contrast, a greater proportion of patients within the age range 18–30 years (55%) were associated with motor vehicle accidents than those aged 31–65 years (30%, p < 0.05) and > 65 years (6.1%, p < 0.001). Similar findings were seen with the analysis of type II and type III fractures with regards to trauma type and injury mechanisms.

Fig. 4
figure 4

Degree of trauma by age groups 18–30 years; 31–65 years and 66 years or older

Full size image

Injury patterns in males and females

A greater proportion of females (89%) sustained fractures in low-energy traumas compared to males (83%, p = 0.01, Supplementary Tables 2a-c). Subgroup comparisons within each type of injury mechanism did not show any sex differences, except for motor vehicle accidents being more prevalent in males (12% vs. females 8.1%, p = 0.04). In type II fractures, no differences were found between sexes in terms of the degree of trauma. Furthermore, an equal number of males and females were diagnosed with type II fractures in response to any of the registered injury mechanisms. Conversely, in type III fractures, males (23%) were more commonly recorded to be involved in a high-energy trauma than females (14%, p = 0.03). Consistent with the total study population, a greater proportion of males with type III fractures (16%) were involved in motor vehicle accidents than females (9.3%, p = 0.04). No other injury mechanism differed between sexes in the type III fracture group.

Discussion

The main findings of this study were that odontoid fractures usually occurred in older individuals, and type II fractures was the most frequent fracture type. Only 14% of the odontoid fractures were due to high-energy trauma. In younger individuals, motor vehicle accidents were usually reported as the injury mechanism. Conversely, in older patients, low-energy traumas were more frequent, and the fractures were predominantly caused by low-level falls. Males were commonly injured in high energy-related accidents. Lastly, non-surgical treatment was the preferred choice of management in 81% of cases.

Data from the SFR demonstrate that odontoid fractures were frequent in elders, with a mean age of 77.2 years and is consistent with previous studies [5, 9]. However, contrary to earlier reports, the present study found a clear unimodal age distribution, with a peak in the seventh to ninth decade of life. In a recent study from Illinois, USA, bimodal age distributions with peaks at 22 and 78 years of age were reported [10]. When comparing traffic accidents between Sweden and Illinois, USA, there were 4.6 times more fatal motor vehicle accidents per capita in Illinois in 2021 [11, 12]. In light of these findings, the peak in younger individuals may be the result of serious traffic accidents which was rather rare in the current study.

Approximately half of the patients with an odontoid fracture were male individuals, which is comparable with earlier studies [2, 4, 9, 13]. Similar to a previous cohort, the included male patients tended to be younger, while elders more commonly tended to be females [5]. Apart from being younger, males were less likely to be injured in low-energy traumas. While the exact mechanisms are not understood, young males are more likely to be risk-takers and show different travelling habits than females. In comparison, elderly females sustain fractures more easily, perhaps because of menopause and subsequently, an osteoporotic odontoid process [14]. Only 10% of the patients in this study had an osteoporosis diagnosis at the time of injury. Previous work has shown rates of osteoporosis in populations with odontoid fractures ranging between 58 and 77% which is considerably higher than this study [15, 16]. Likely explanations to this discrepancy could be that patients in this study possibly demonstrated signs of osteoporosis without having the diagnosis recorded (ICD codes), or that it was undiagnosed.

A large majority of patients had an intact neurological status (94%) whereas only 2.1% had some type of spinal cord injury. This is somewhat lower than previous reports with spinal cord injuries affecting 3–10% of patients [5, 17,18,19,20]. The relatively low rate in this study could be explained by several morphological attributes. For instance, the relatively large cross-sectional diameter of the spinal canal at the level of the odontoid process compared to the diameter of the spinal cord may result in fewer spinal cord injuries compared to other cervical spine fractures. Moreover, most odontoid fractures were associated with low-energy trauma and therefore did not contribute to large displacement causing neurological injuries.

The occurrence of odontoid fractures varied during the seasons with peaks in May and August. There is, to our knowledge, no previous research on the seasonal variation of odontoid fractures. However, other studies from the SFR have investigated seasonal variation of various fracture types. Talus fractures were usually sustained in warmer months (May through October) [21] while patella and ankle fractures in the colder months (October through March) [22, 23]. Speculatively, individuals are more physically active and spend more time outdoors during the summer months, therefore falling more often. Moreover, higher temperatures are associated with hypovolemia, syncope and other cardiac events in the elderly causing falls, and ultimately, odontoid fractures. In contrast, in the winter months, the streets are often slippery in Sweden which could be the cause of the increased number of patella and ankle fractures. One could only speculate as to why this increase was not seen in odontoid fractures. Perhaps those with odontoid fractures are generally older with more co-morbidities than those sustaining patella or ankle fractures. This could result in patients remaining inside when they know that the ground is unsafe to walk on.

Most of the included patients were treated non-surgically (81%) which is consistent with other publications [20, 24]. However, a previous study based on a Swedish population reported that 53% with a type II odontoid fracture underwent surgery with an annual increase between 2002 and 2014 [25]. This study included urban hospitals whereas the SFR receives data from hospitals all over Sweden. Moreover, the SFR has only collected data on spine fractures since 2015 which could provide another explanation to the different managements of type II fractures. Type II fractures are generally prone to pseudoarthrosis and non-union, and although guidelines on the management of odontoid fractures are absent, the risk of impaired fracture healing in type II fractures is an indication for surgical stabilization. This is consistent with our study since significantly more patients with type II fractures underwent surgery than the other fracture types. The overall aim of surgical treatment is to preserve the neurological function, relieve pain and establish a stable fusion. As only 3.5% of the included patients suffered neurological deficits, this could partly explain why most of the patients received non-surgical treatment. Other possible reasons may include high age and major co-morbidities.

Lastly, in this study, 87% of the fractures were caused by falls. For type II fractures, the corresponding number was 90% and for type III fractures 85%, respectively. This is similar to previous publications from Scandinavia [5, 17]. Consistent with earlier findings, older patients were more often injured after falls from a standing height or less, while younger patients were more often injured in a motor vehicle accident [10, 13]. Furthermore, older patients were more likely to sustain an odontoid fracture by a low-energy trauma than younger patients which could be explained by altered body composition, neurological dysfunction and the increasing number of co-morbidities that parallel with aging. For instance, physical inactivity, impaired balance, and neurological disabilities may result in unintended falls. In addition, polypharmacy is common in elders, further increasing the risk of falls [5]. Others have proposed that increasing age is associated with a stiffer cervical spine, and a reduced bone density, which can explain why elderly need less energy for a fracture to develop [25].

Strengths and limitations

The SFR contains data from patients within various geographic locations, representing a heterogenous population which enables a broad generalizability of the results from this study. Additionally, compared to traditional epidemiologic studies based on a sample of the population, the SFR is a population-based register minimizing the risk of selection bias. Moreover, the SFR contains one of the largest datasets on odontoid fractures, enabling the conduction of reliable subgroup analyses. The SFR also contains detailed records pertinent to injury patterns and the data are prospectively collected, which mitigates the risk of recall bias. On the other hand, the completeness of odontoid fracture coverage in the SFR is unknown. Fractures are registered by physicians with various levels of expertise, which may introduce bias due to incorrectly diagnosed or classified fractures. While the reliability of the Anderson and D’Alonzo classification [8] in the SFR has not been evaluated, classification of other fractures including thoracolumbar [26], ankle [27], and humerus [28] fractures have been shown to be highly reliable.

Conclusion

Based on the SFR, the typical patient with an odontoid fracture was older and suffered a type II fracture. Moreover, odontoid fractures typically occurred as a result of a low-energy trauma with a relatively low risk of neurological deficits. However, younger patients and males were more commonly involved in motor vehicle accidents. Most patients with an odontoid fracture were treated non-surgically although surgical treatment was more frequent in type II fractures.