Introduction

The rapidly aging population in the Western world presents a significant healthcare challenge, as this demographic is particularly susceptible to fractures and will undoubtedly play a pivotal role in shaping the future of medical care. According to the Global Burden of Disease Study females had an age-standardized rate of 92.2 vertebral fractures per 100,000 people, while males had a rate of 125.3 per 100,000 in Central Europe [1]. Recently, a staggering 94% increase in cervical spine fractures in Germany between 2009 and 2019 was demonstrated, with a notable shift in the age distribution of affected individuals, and patients aged 70 or older accounting for 70% of all cases [2]. In fact, axis fractures had an incidence rate of 8.2 per 100,000 inhabitants in Germany in 2019, with an 82.5% increase since 2009, and predominantly affected males and geriatric patients aged 80 years or older, accounting for 59.1% of cases [3]. The most common type of axis fracture of patients aged 80 years or older is the dens axis (odontoid) fracture, with type II fractures according to Anderson and D'Alonzo being the most prevalent, followed by type III fractures [4,5,6,7,8]. The current epidemiological trend has inspired a plethora of retrospective studies, reviews, and meta-analyses in the literature investigating the morbidity and mortality of geriatric patients. The optimal treatment strategy is an ongoing controversy and the discussion has been shifted in recent years primarily to patients older than 80 years [9,10,11]. While non-surgical management with external immobilization has been the traditional approach for type I and some type II fractures, recent evidence suggested that surgical intervention may provide better outcomes, particularly for unstable type II and type III fractures [10, 12]. However, surgery may be associated with higher risks of complications, such as infection and neurovascular injury, and may not be suitable for all patients, especially if “modifiers” like osteoporosis, a nonreducible fracture or a reverse fracture line among others are present [12, 13]. Therefore, a careful evaluation of the fracture type, patient factors, and potential risks and benefits of each treatment approach should be considered before making a decision [13, 14]. Despite the availability of up-to-date treatment guidelines and research efforts, the mortality of odontoid fractures in geriatric patients is still high and according to Venkatesan et al., it is even higher than hip joint-related fractures [9, 15]. Graffeo et al. recently suggested, that among octogenarian patients with type II odontoid fractures, frailty was associated with increased mortality, independent of the chosen treatment strategy [9]. Despite several meta-analyses that provide several risk factors for in-hospital mortality in geriatric patients with axis fractures, the general conclusions of most studies are limited by the low quality of evidence available, highlighting the need for randomized controlled trials and large register-based analyses [12, 16, 17].

To provide data with a high degree of generalizability we conducted a nationwide cross-sectional study, aiming (1) to identify risk factors for high in-hospital mortality in patients aged 80 and older with axis fractures. (2) The second aim was to identify factors that were associated with in-hospital mortality in the surgically treated patient cohort.

Methods

We conducted a cross-sectional nationwide study with pooled data from 2019 through 2021. In accordance with Section 17b of the German Hospital Financing Act, a universal, performance-based, and flat-rate remuneration system has been introduced for general hospital services. The basis for this is the German Diagnosis Related Groups system (G-DRG system), whereby each inpatient case of treatment is remunerated using a corresponding DRG lump sum payment. The Institut für das Entgeltsystem im Krankenhaus (InEK) GmbH provides detailed data on the main diagnoses based on “International Statistical Classification of Diseases and Related Health Problems, 10th Revision”  (ICD-10) codes, age- and sex distribution, length of hospital stays, reasons for discharge (including “death”), coded concomitant diagnoses (based on ICD-10 codes) and coded procedures based on “Operationen- und Prozedurenschlüssel” (OPS) codes. The InEK browser enables analysis back to the year 2019. For the following comprehensive analysis, all cases from 2019 through 2021 were extracted and pooled. Based on the ICD-10 code for axis fracture (S12.1), data for total case numbers, the numbers of coded concomitant diagnoses, and the number of in-hospital deaths were extracted. In the second step, case combinations of the ICD-10 code S12.1 with any surgical procedure of the spine (OPS-codees 5–83.-; 5.030.-) for the years 2019 through 2021 were extracted and pooled to conduct the same analysis as mentioned above for surgically treated patients exclusively.

The study included only patients aged 80 years or older, for these have been in the recent focus of debate regarding the optimal treatment. The analysis was conducted on comorbidities that were documented in at least 1.0% of cases. The primary outcome parameter was the event of in-hospital death. To identify potential risk factors for the in-hospital mortality, the prevalence of comorbidities, concomitant injuries (each by ICD-10 code), and procedures (by OPS-codes) were compared between two groups: Group A (Survival Group) and Group B (Deceased Group). In the next step, cases with documented surgical procedures were further divided into these two groups (A and B) to investigate potential risk and protective factors for the in-hospital mortality in surgically treated patients.

Ethics statement

Informed Consent and Investigational Review Board (IRB) were not required for this cross-sectional study as it used data from an anonymous, de-identified, administrative database.

Statistical analysis

Data are presented as total numbers and as percentages of the total. The Chi-Square test was used for statistical comparison between groups (A and B). To analyze the association between potential risk factors and the outcome parameter of in-hospital mortality the odds ratios (OR) with 95% confidence intervals (95% CI) were calculated for the previously identified factors.

Results

Patient characteristics

In total, N = 10,077 hospital admissions cases were pooled from 2019 to 2021 and analyzed. Only patients aged 80 years or older were included. In total n = 843 cases of in-hospital deaths were documented, giving an in-hospital mortality rate of 8.4%. The proportion of male patients was 35.2% (n = 3544) and 64.8% (n = 6533) for female patients (p < 0.001). In the deceased group, 47.7% (n = 402) were male and 52.3% (n = 441) were female patients. Among all cases, n = 201 (2.0%) were classified as short-stay patients, n = 8876 (88.1%) as normal-stay patients, and n = 1000 (9.9%) as long-stay patients. Ten-point seven percent of cases (n = 1076) had severe motor function impairment (Barthel Index: 20–55 points). Mild to moderate motor function impairment (Barthel Index: 60–75) was documented in 4.1% of cases (n = 417) and 1.4% of cases (n = 142) had mild motor function impairment (Barthel Index: 80–95 points). A significantly higher proportion of cases were classified as having ´very severe´ motor function impairment (Barthel Index: 0–15 points), whereas significantly fewer cases with ´mild,´ ´moderate,´ and ´moderately severe´ motor function impairment were documented in the deceased group (all p < 0.01; Table 1).

Table 1 Comparison of the prevalence of comorbidities in all patients with axis fractures and the deceased patient cohort
Full size table

Risk factors for in-hospital mortality

The most prevalent comorbidity was essential hypertension, which was reported in n = 6731 cases (67.3%). Codes for essential hypertension were significantly more often used in the survival group than in the deceased group (p < 0.001). In only n = 34 cases (0.3%) obesity and in n = 1665 (16.5%) cases a form of Type 2 Diabetes Mellitus was coded with no difference between the two groups (both p > 0.05). A code for atrial fibrillation was used in n = 2837 (28.2%) of cases and the proportion was significantly higher in the deceased group (n = 315; 37.4%; p < 0.001). Further, left-sided heart failure was significantly more prevalent in the deceased group (n = 201; 23.8% vs. n = 1024; 11.1%; p < 0.001). The frequencies of codes for volume deficiency (n = 107; 12.7%), hyperkalemia (n = 59; 7.0%), hypokalemia (n = 197; 23.4%), acute kidney failure (n = 148; 17.6%) and chronic kidney disease (n = 257; 30.5%) were significantly higher in the deceased group (all p < 0.01). Hospital-acquired pneumonia was documented in n = 444 (4.4%) of all cases and significantly more frequently in the deceased group (n = 133; 15.8%; p < 0.001). There was a statistically significant difference between the groups in the prevalence of acute anemia (due to hemorrhage) which was documented in the deceased group in n = 193 (21.3%) cases (p < 0.001). Similarly, a significantly higher portion of cases of hemorrhagic diathesis was found in the deceased group (n = 78; 9.3% vs. n = 249; 2.7%; p < 0.001). However, no difference in the frequency of long-term anticoagulant therapy was found between the groups (p = 0.536). The prevalence was 25.6% in all cases. Most codes for cognitive disorders and motor dysfunctions (according to the Barthel Index stratification) were more prevalent in the deceased group. Table 1 provides an overview of all analyzed comorbidities in the total cohort (A+B) and the deceased group (B).

The most common concomitant injuries were head and face wounds in 25.9% of all cases, followed by a concussion in 12.8% of all cases. Traumatic brain injuries were rare, traumatic subdural hemorrhage was documented in n = 161 (1.6%), and traumatic subarachnoidal hemorrhage in n = 137 cases (1.4%). Traumatic subdural hemorrhage was significantly more prevalent in the deceased group (n = 29; 3.4%; p < 0.001). The frequency of Cranium and facial skeleton fractures was documented in n = 945 (9.4%) cases and was significantly more frequent in the deceased group (n = 127; 15.1%; p < 0.001). Concomitant cervical fractures were documented in 11.6% of all cases represented by 1st cervical vertebra (atlas) fractures and 7.0% of all cases represented by subaxial cervical spine fractures. Atlas fractures were significantly more frequent in the deceased group compared to the survival group (n = 147; 17.4% vs. n = 1020; 11.0%; p < 0.001). There was a significant difference in the frequency of fractures of the rib cage which was documented in n = 80 (9.5%) cases of the deceased and n = 324 (3.5%) of the survival group (p < 0.001). Cervical Spinal cord injuries (SCI) were rare, but significantly more frequent in the group of deceased patients (n = 64; 0.7% in group A vs. n = 17; 2.0% in group B; p < 0.001). In Table 2, all analyzed concomitant injuries in the total cohort and the deceased group are listed.

Table 2 Comparison of the prevalence of concomitant injuries in all patients with axis fractures and the deceased patients cohort
Full size table

Surgical procedures

A total of n = 3178 (31.5%) patients in this cohort were treated surgically. The in-hospital mortality rate in the surgically treated patients was 9.4% (n = 298 cases with in-hospital deaths vs. n = 545; 7.9% in the non-surgically treated group). The OR for in-hospital mortality depending on the surgical treatment was 1.13 (95% CI: 0.99–1.3; p = 0.103). In n = 1954 (61.5%) cases a dorsal approach was used and in n = 1266 (39.8%) cases a ventral approach was, which indicates, that a 360° approach was applied in n = 42 (1.3%) cases. Most frequently, screw fixation for one segment was applied (n = 1404 cases, 44.2%) followed by fixation with a screw-rod system for 1 segment (n = 919 cases; 28.9%). Multisegmented fixation with a screw-rod system of up to 5 segments was conducted in n = 728 (22.9%) cases. A dorsal fusion of 1–5 segments was coded in n = 823 (25.9%) cases. Decompression at one segment was only documented in n = 62 (2.0%) cases (Table 3).

Table 3 Frequencies of surgical procedures by documented OPS-codes in n = 3178 surgically treated patients
Full size table

Risk factors in surgically treated patients

In the group of surgically treated patients (n = 3178) the prevalence of hospital-acquired pneumonia had a high OR for in-hospital mortality among comorbidities, with 2.12 (95% CI: 1.60–2.80; p < 0.001). Hemorrhagic diathesis due to medication and congestive heart failure also had high ORs, with 1.98 (95% CI: 1.42–2.77) and 1.80 (95% CI: 1.37–2.37), respectively (both p < 0.001). Chronic kidney disease stage 3 or higher had a modestly elevated OR of 1.44 (95% CI: 1.09–1.89; p = 0.02). The other evaluated conditions, including dementia, diabetes, urinary tract infections, and coronary heart disease, showed ORs close to 1, indicating little to no association with in-hospital mortality (Fig. 1).

Fig. 1
figure 1

Risk factors for in-hospital mortality in surgically treated patients: ORs for comorbidities are displayed with 95% CI and results of the Chi-squared test. Whiskers indicate the 95% CI. OR = 1.00 is indicated by the dotted vertical line

Full size image

Considering the concomitant injuries and factors, we found the odds for the in-hospital mortality to be the highest in patients with SCI, although only prevalent in 1.1% (n = 34) of surgically treated patients (OR = 4.62, 95% CI: 2.23–9.58; p < 0.001). Further, acute renal failure (OR = 3.20, 95% CI: 2.26–4.53, p = 0.000) and acute bleeding anemia (OR = 2.06, 95% CI: 1.64–2.59, p = 0.000) were associated with a higher in-hospital mortality rate compared to cases without these conditions. Traumatic subdural hemorrhage (OR = 2.04, 95% CI: 1.16–3.61, p = 0.080) and concomitant atlas fracture (OR = 1.34, 95% CI: 1.00–1.80, p = 0.147) were not associated with statistically significant higher in-hospital mortality odds (Fig. 2).

Fig. 2
figure 2

Risk factors for in-hospital mortality in surgically treated patients: ORs for concomitant diseases, injuries and complications are displayed with 95% CI and results of the Chi-squared test. Whiskers indicate the 95% CI. OR = 1.00 is indicated by the dotted vertical line

Full size image

The cases in which blood transfusion was indicated were associated with a significantly, 2.02-fold increased OR of in-hospital mortality (95% CI: 1.60–2.55, p = 0.000). Further, multi-segmental instrumentation using a screw-rod system with 4 or more segments was by tendency associated with 1.79-fold increased odds of in-hospital mortality (95% CI: 1.15–2.79, p = 0.052).

As seen in Fig. 3, most surgical techniques were not associated with a relevant change in the OR for in-hospital mortality. However, patients who were dependent on care level 3 had 1.39-fold increased odds (95% CI: 1.06–1.83, p = 0.035). Screw fixation for 1 segment as applied for anterior dens-screws was not associated with a significantly elevated OR of in-hospital mortality (OR 1.14; 95% CI: 0.92–1.40; p = 0.260). On the other hand, fixation by a screw-rod system of one segment, as used for encoding C1/C2 fusion procedures (e.g., Goel-Harms technique) showed a moderate negative association with in-hospital mortality (OR = 0.74, 95% CI: 0.56–0.97; p = 0.044).

Fig. 3
figure 3

Risk factors for in-hospital mortality in surgically treated patients: ORs for surgical procedures and interventions are displayed with 95% CI and results of the Chi-squared test. Whiskers indicate the 95%CI. OR = 1.00 is indicated by the dotted vertical line

Full size image

The use of intraoperative navigation was found to have an OR of 0.45 (95% CI: 0.16–0.71, p = 0.001) for in-hospital mortality. Figure 3 lists the ORs for in-hospital mortality depending on all analyzed procedures (by OPS code) for surgically treated patients.

Discussion

This nationwide cross-sectional register study on 10,077 geriatric patients aged 80 or older, that suffered axis fractures revealed an association between certain comorbidities and concomitant injuries with in-hospital mortality. In surgically treated patients, comorbidities played a less central role, but hospital-acquired pneumonia and acute renal failure are associated with a more than twofold OR.

General risk factors for in-hospital mortality in geriatric patients with axis fractures

Comorbidities indicating frailty and severe illness were associated with higher in-hospital mortality, such as essential hypertension, atrial fibrillation, left-sided heart failure, chronic kidney disease, dementia, and very severe motor dysfunction (Barthel index: 10–15 pts.). These findings are consistent with previous studies that have shown the importance of comorbidities and concomitant injuries in predicting mortality in elderly patients with axis fractures [6, 18]. In general, comorbidities are known to be associated with elevated risk for mortality after cervical spine fractures [19]. Concordant with our findings, chronic kidney disease is one of the key prognostic factors for early mortality in older patients with traumatic cervical spine injuries [20]. Shafafy et al. identified several predictors of mortality in elderly patients with fractures of the odontoid process [21]. They found that head injury and the presence of other spinal injuries increase the risk of mortality after 30 days [21]. Similarly, we revealed that concomitant injuriese such as SCI, traumatic brain injuries, atlas fractures, subaxial vertebral fractures, and rib cage fractures were indicators for high-risk in-hospital mortality. The knowledge of these factors can help to identify patients at risk, which should be monitored very carefully. Anyhow, there was no statistically significant difference in mortality rates between surgically treated (9.4%) and non-surgically treated patients (7.9%; p = 0.103) in the current cohort. Smith et al. retrospectively analyzed the complications of surgical versus conservative treatment of isolated type II odontoid fractures in octogenarians and found, that no relevant difference in the acute in-hospital mortality rate between 15.0% in the nonsurgical group and 12.5% in the surgical group [22]. These findings are in line with other retrospective studies [23]. Similarly, a meta-analysis by Deng et al. on the treatment of odontoid fractures in the octogenarians on 22 cases series and retrospective studies including 248 patients reports on a higher mortality rate of 10.1% for non-surgery vs. 5.4% for surgery, but without statistical significance [12]. Yang et al. conducted a meta-analysis to grade the evidence on conservative versus surgical treatment for type II odontoid fractures in the elderly. Analyzing twelve studies involving 730 patients aged 60 years or older, they did not find a difference with regard to the mortality between the two procedures [16]. In a prospective study on 336 patients, Rizvi et al. identified major comorbidities and older age as significant factors contributing to physicians’ decision to refrain from the surgical fixation of odontoid fractures [6]. Prospective trials are emerging to better understand the outcomes associated with different treatment options. Currently Robinson et al. are conducting the Uppsala Study on Odontoid Fracture Treatment (USOFT) which is one of the first randomized controlled trials comparing non-surgical and surgical management of type II odontoid fractures in the elderly [24]. Regarding stability and fusion rates meta-analysis see an advantage for the surgical treatment [12].

Risk factors for the in-hospital mortality in surgically treated patients

In the current cohort 31.5% of the patients were treated surgically. Most frequently, a screw fixation for one segment was applied as used to encode for dens-screws or transarticular “Magerl” screws, followed by fixation with a screw-rod system for 1 segment as used to encode the dorsal Atlanto-axial fusion techniques. The role of comorbidities was less pronounced in this cohort of surgically treated patients. We revealed Pneumonia to be a relevant risk factor (OR = 2.12). The occurrence of pneumonia after surgery is a frequent complication that is linked with significant morbidity and mortality [25]. Aspiration may be partially accountable, which is associated with significantly greater length of stay, costs, and mortality and has been reported with an incidence of 5.3 cases of per 1000 cervical procedures [26]. The highest OR in this sudy of 4.62 for in-hospital mortality in surgically treated patients was associated with SCI, although it was only prevalent in 1.1% of the cases. SCI has been well recognized as a significant factor for a fatal outcome in cervical spine fractures, especially in the geriatric population [27, 28]. However, low case numbers of SCI in this cohort prevented statistical differentiation between complete and incomplete SCI. Furthermore, only SCI diagnoses coded as concomitant but not primary ICD-10 code could be analyzed. Acute renal failure was found to have an OR of 3.20 for mortality in our cohort and could potentially be suggested as another significant predictor. Acute renal failure is well-known to be associated with increased mortality and therefore should be meticulously screened for and treated as early as possible [29, 30]. The current analysis suggests that bleeding complications are likely to play a major role for the in-hospital mortality, with medication-induced hemorrhagic diathesis having a high OR of 1.98. The deceased group had a significantly higher prevalence of acute anemia due to hemorrhage and hemorrhagic diathesis compared with the survival group. The cases in which blood transfusion was indicated were associated with 2.02-fold increased odds of in-hospital mortality. However, no difference was found in the frequency of long-term anticoagulant therapy between the groups, which was present in 25.6% of all cases. It has been demonstrated, that massive intraoperative hemorrhage is an important risk factor for in-hospital mortality after spine surgery with a reported OR of 28.2 [31]. For planned spine surgery various guidelines on the perioperative management of anticoagulation have been published [32, 33]. The initial findings from the AO Spine Anticoagulation Global Survey recognized the challenge of offering consistent perioperative anticoagulation advice to patients due to the lack of universally acknowledged literature or guidelines [34]. Cutler et al. demonstrated, that next to age and white race the history of bleeding disorders were independently predictive of complications in the multivariate analysis in their retrospective cohort of 103 patients treated with anterior fixation for odontoid fractures with a significant OR of 4.4 [35]. Our results and the body of the literature suggest that investigating the peri-operative management of coagulopathies to mitigate adverse events and exploring the feasibility of bridging anticoagulants in the trauma setting should be further considered. Hemoglobin levels might be employed to identify high-risk patients for in-hospital mortality and could be integrated into a predictive score system [21].

The geriatric population is prone to an increased risk of complications during the peri-operative period caused by prevalent comorbidities and weak bone structure. Nonetheless, Robinson et al. propose, that surgical treatment improves survival of elderly with axis fracture based on their national population-based multi-registry cohort study on 3375 patients [37]. A systematic analysis of the literature on the morbidity and mortality related to odontoid fracture surgery in the elderly population showed that major complications after surgery include cardiac failure, deep vein thrombosis, stroke, pneumonia, respiratory failure, liver failure, and severe infection [38]. Notable, anterior surgery had a higher rate of site-specific complications such as nonunion and revision surgery [38]. Longo et al. report a mortality rate of 7.6% and a complication rate of 9.1% in elderly patients treated with anterior odontoid screw fixation and described functional dependency as one predictor for mortality, which is in line with our findings [36]. Interestingly, we identified the fixation using a screw-rod system for one segment (e.g., as in "Goel/Harms" technique) to be associated with a significantly decreased risk for in-hospital mortality (OR = 0.74). Based on the results of their retrospective study of 70 patients above 75 years of age with odontoid fractures Faure et al. conclude, that the Harms technique showed lower risk of complications and better mechanical stability compared to anterior screw fixation [39]. In line, Cutler et al. report on a high complication rate (37.9%) and relevant mortality rate of 6.8% of anterior fixation of odontoid fractures in a retrospective analysis of 103 patients aged 73.9 years at average [35].

One further protective surgical factor was the use of intraoperative navigation. Due to substantial improvements and rapidly increasing availability, computer-assisted navigation is establishing itself as the gold standard for instrumentation at the cervical spine. It has been reviewed to increases the accuracy of pedicle screw placement, to reduce screw mispositioning and leading to fewer revision surgeries [40]. Ishak et al. retrospectively analyzed 35 patients suffering type II odontoid fractures with an average age of 86.5 years, that were treated with a navigated C1-C2 fusion. They found no in-hospital deaths and the overall major complication rate was 11%, leading to their conclusion, that Atlanto-axial fusion by using intra-operative spinal navigation is a safe and effective procedure [41].

Limitations

The findings of this study should be interpreted within the scope of its limitations. In general, registry studies are commonly not accepted by design to allow causative relationships to be made. The ICD-10 coding system does not allow for the differentiation of axis fractures (ICD10: S12.1) into odontoid or corpus or vertebral arch fractures. Furthermore, the InEK database does not provide any information about the fracture classification or dislocation. Thus, this study is based on the assumption, that the majority of axis fractures in the geriatric population are odontoid fractures. Further, the database employed did not enable to analyze the mid-to long-term results, but only in-hospital mortality. However, while focusing on risk factors in surgically treated patients this analysis adds important information to be considered for the indication for surgery. The retrospective character of this database study is prone to selection bias. The main statistical limitation is present in the analysis of frequency data, as it does not allow for the consideration of multiple variables affecting the outcome variable in-hospital mortality. Additionally, internal validity checks, such as bootstrapping or cross-validated estimates cannot be performed with this frequency data. The generalizability of this study's results may be limited due to its focus on German patients aged 80 years or older and the retrospective design. Differences in healthcare systems, clinical practices, and patient populations could impact the applicability of the findings to other settings or age groups. Therefore, it is important to carefully interpret the data and consider alternative approaches, such as prospective studies to gain a deeper understanding of the association between the tested risk factors and in-hospital mortality.

Conclusion

This study provides important insights into the factors associated with in-hospital mortality in geriatric patients aged 80 and older with axis fractures, which can aid in the identification of high-risk patients. The findings underscore the importance of comprehensive geriatric assessment and optimization of comorbidities prior and during treatment. The indication for surgical treatment must be carefully individualized, respecting the comorbidities and concomitant injuries and tailored treatment should be aimed for. We suggest, that prospective, multicentric studies should focus on the choice of the surgical technique, perioperative blood management and intraoperative navigation as potential protective factors to further improve the standard of care in the geriatric population.