Predictors of In-Hospital Mortality in Patients with New Diagnosis of Type II Odontoid Fractures: A Retrospective Level I Trauma Center Experience

Introduction: Type II Anderson-D’Alonzo odontoid fractures are common and may present in isolation or in the setting of multi-trauma. This study sought to identify independent predictors of in-hospital mortality and post-operative length of stay in patients with new diagnosis of Type II odontoid fracture.

Methods: From an initial cohort of 457 patients admitted to a Level 1 Trauma Center between January 2016 and March 2024, 237 were confirmed by a neurosurgeon to have newly diagnosed Type II odontoid fractures on imaging. After checking for collinearity, significant variables from univariate analysis were included in a logistic multivariable model for in-hospital mortality status. A linear multivariate model was used to identify independent predictors of post-operative length of stay following univariate analysis.

Results: Mean age was 75.5 (SD 6.4) years, and 54% of patients were male. Fall was the most common injury mechanism (81.7%), followed by motor vehicle collision (12.3%), assault (3.4%), and pedestrian incidents (2.6%). Multiple trauma was present in 22.33%. Mean Injury Severity Score (ISS) was 12.77 (SD 7.1), and mean GCS was 13.95 (SD 2.75). Twenty-four patients (10.1%) had traumatic brain injury (TBI) and 14 patients (5.9%) had spinal cord injury (SCI). Overall mortality was 9.7%. We observed a bimodal distribution of mortality by age, with peaks at 25-35 years and 65-95 years. Multivariable logistic analysis demonstrated that older age (p=0.01), ISS (p-0.034), pulse rate (p=0.01), and presence of SCI (p=0.004) were independent and significant predictors of in-hospital mortality. In the seventy-eight patients who underwent posterior fusion, multivariate linear analysis identified GCS (p < 0.001), admission Revised Trauma Score (RTS) (p=0.011) and presence of SCI (p < 0.001) as independent predictors of increasing post-operative length of stay (LOS).

Conclusion : There was a bi-modal age distribution of in-hospital mortality among patients admitted with new diagnosis of Type II odontoid fracture. Higher ISS and presence of SCI but not TBI independently predicted in-hospital mortality. Higher admission heart rate also independently predicted in-hospital mortality. Except for age, demographic variables including 5-item Frailty Index, Area Deprivation Index, and smoking status did not predict in-hospital mortality. Lower GCS, higher admission RTS, and presence of SCI were independent predictors of post-operative LOS.