Fracture morphology-driven surgical strategy in ankylosing spondylitis: when does sagittal imbalance mandate pedicle subtraction osteotomy?

BACKGROUND

Thoracolumbar fractures in patients with ankylosing spondylitis (AS) pose unique challenges due to altered spinal biomechanics and sagittal imbalance. Optimal surgical strategies remain controversial, and objective criteria for selecting between deformity-corrective and fusion-only procedures are lacking.

METHODS

We retrospectively reviewed 41 AS patients with thoracolumbar fractures who underwent surgical treatment. Patients were categorized as vertebral body (VB) type or intervertebral space (IS) type fractures. The primary outcome was surgical strategy selection, defined as pedicle subtraction osteotomy (PSO) versus non-PSO procedures. Propensity score matching (PSM) was performed using preoperative clinical and radiographic parameters to minimize selection bias in comparing outcomes between fracture types. Radiographic outcomes, pain improvement, and complication rates were compared. Predictive modeling using logistic regression and decision tree analysis was conducted to identify determinants of surgical strategy.

RESULTS

A total of 41 patients (24 VB-type, 17 IS-type) were included. The VB group exhibited significantly greater preoperative sagittal imbalance and underwent PSO more frequently, while IS-type fractures were treated with posterior spinal fusion (PSF). PSM analysis (n = 17 per group) confirmed greater radiographic correction in the VB group without differences in pain outcomes. The decision tree model achieved an accuracy of 82.9% and AUC = 0.88 (95% CI: 0.85–0.99), with C7SVA > 222.6 mm as the primary decision criterion. Multivariable logistic regression identified C7SVA as the only statistically significant predictor (OR = 1.013; 95% CI: 1.001–1.026; p = 0.039). These findings suggest promising but preliminary discriminative ability, and external validation is warranted.

CONCLUSIONS

In AS patients with VB-type thoracolumbar fractures, preoperative C7SVA > 222.6 mm represents a data-supported threshold for PSO, while C7SVA > 186.8 mm combined with BMD ≤ − 2.25 identifies a secondary risk group warranting PSO consideration. PSF alone appears sufficient for IS-type injuries with relatively preserved sagittal alignment. Preoperative sagittal profiling and bone mineral density assessment enable individualized, data-driven surgical planning in this high-risk population.