Radiographical Predictors for Postoperative Sagittal Imbalance in Patients With Thoracolumbar Kyphosis Secondary to Ankylosing Spondylitis After Lumbar Pedicle Subtraction Osteotomy

Study Design. A retrospective radiographical study. Objective. To identify the radiographical predictors for sagittal imbalance in patients with thoracolumbar kyphosis secondary to ankylosing spondylitis (AS) after 1-level lumbar pedicle subtraction osteotomy (PSO). Summary of Background Data. Few studies had correlated the preoperative sagittal parameters with postoperative sagittal alignments to determine the radiographical predictors for postoperative sagittal imbalance in patients with AS after 1-level lumbar PSO. Methods. Thirty-six patients with thoracolumbar kyphosis secondary to AS who underwent 1-level lumbar PSO were recruited with a minimal follow-up of 24 months (mean = 27.4 mo; range, 24-53 mo). Correlation analysis and subsequent stepwise multiple regression analysis were used to evaluate the correlations between preoperative parameters, including global kyphosis, local kyphosis, thoracic kyphosis, thoracolumbar Cobb angle, lumbar lordosis, pelvic incidence (PI), pelvic tilt, sacral slope, and sagittal vertical axis (SVA), as well as SVA at the last follow-up. All these patients were further divided into 2 groups according to the PI value (group A: PI >50 degrees; group B: PI <= 50 degrees). The correction outcomes were compared between these 2 groups. Results. The preoperative SVA was not significantly different between group A and group B (157.6 mm vs. 124.5 mm; P > 0.05), and both groups had similar magnitudes of kyphosis corrections at the last follow-up (global kyphosis: 42.9 degrees vs. 46.1 degrees; local kyphosis: 42.7 degrees vs. 40.5 degrees; lumbar lordosis: 35.7 degrees vs. 43.0 degrees). However, group A patients had significantly larger SVA at the last follow-up (73.2 mm vs. 28.7 mm; P < 0.05) and a higher incidence of postoperative sagittal imbalance (77.8% vs. 25.9%; P < 0.05) than those in group B. The stepwise multiple regression analysis demonstrated that both preoperative SVA and PI were significant independent predictors of postoperative sagittal alignments, which explained 52.0% and 9.7% of the variability of SVA at the last follow-up, respectively. Conclusion. Patients with AS with either larger preoperative SVA or larger PI are more likely to experience failed sagittal realignments after 1-level lumbar PSO. For these patients, additional osteotomies may be recommended for satisfactory correction outcomes.