9 - Outcomes Following Non-operative Management of Three-column Fractures in the Ankylosed Spine

Introduction:
Spinal fusion remains the gold standard management for three-column fractures. Spinal ankylosis seen in ankylosing spondylitis (AS) and diffuse idiopathic skeletal hyperostosis (DISH) results in increased brittleness and altered spinal biomechanics as well as increased susceptibility to unstable injury and neurologic deficit even with low energy trauma. The high medical comorbidity of elderly individuals with AS and DISH may lead to disproportionate operative risk even in individuals with unstable fracture patterns.

Aims/

Objectives: This study investigates the relative efficacy of non-operative treatment, in patients unfit for surgery for the treatment of three-column fractures of the ankylosed spine, with relative comparison of outcomes to patients undergoing fusion.

Methods:
All patients over age 18 with AS or DISH who sustained traumatic injuries resulting in three-column fractures throughout the cervical, thoracic, and/or lumbar spine at a single academic institution from 1999-2020 were retrospectively identified. Patients with infection or malignancy were excluded. Patient demographics, medical comorbidities, mechanism of injury, fracture displacement, and injury severity scores (ISS) were collected. The decision to proceed with operative vs non-operative treatment was at the discretion of the lead surgeon and reflects shared patient decision making. Operative management included fusion with or without decompression and non-operative management included bracing or extension casting. The primary outcomes assessed were mortality (90-day and one-year post-injury) and neurologic status (ASIA grade) at final clinical follow-up. Secondary outcome metrics included thrombotic events (deep venous thrombosis and pulmonary embolism) and respiratory complications (pneumonia and respiratory failure). Univariate analysis compared primary and secondary outcomes across operative and non-operative treatment groups. A stepwise logistic regression model identified predictors of one-year post-injury mortality.

Results:
A cohort of 138 patients with 153 three-column fractures was identified: 56% with AS, 29% with DISH, and 16% with both diagnoses. When comparing treatment groups (operative, n=116 vs non-operative, n=37), the non-operative group was older (p < 0.001), had a greater Charlson Comorbidity Index (CCI) (p < 0.001), had a lower body mass index (p=0.04), had a greater portion of patients with a DISH diagnosis (p=0.009), and were more likely have sustained a ground-level fall (p=0.02) or have a non-displaced fracture pattern (p=0.04) (Table 1). In univariate analysis, non-operative treatment was associated with increased 90-day (p < 0.001) and one-year (p < 0.001) post-injury mortality. Neurologic status, thrombotic events, and respiratory complications did not differ significantly by treatment (all p>0.05) (Table 2). In logistic regression, CCI (odds ratio [OR]: 1.34, p=0.019), DISH diagnosis (OR: 5.89, p=0.005) and non-thoracic injury level (thoracic OR: 0.29, p=0.026) were significant predictors of one-year mortality. Treatment modality was not a significant predictor of one-year mortality (p=0.80).

Conclusion:
In AS or DISH patients who are not surgical candidates, non-operative management of three-column fractures relative to standard-of-care operative fusion is not associated with short-term neurologic, thrombotic, or respiratory complications. After controlling for confounding factors, non-operative management is not a predictor of mortality.