Validation of Sagittal Age-adjusted Score in Predicting Proximal Junctional Kyphosis/Failure and Clinical Outcomes Following Adult Spinal Deformity Surgery

Objectives. To validate the sagittal age-adjusted score (SAAS) in predicting proximal junctional kyphosis/failure (PJK/F) and good clinical outcomes following adult spinal deformity (ASD) surgery. Summary of Background Data. SAAS is a relatively new assessment system that incorporates age-adjusted sagittal parameters of pelvic incidence (PI)-lumbar lordosis (LL), pelvic tilt (PT), and T1 pelvic angle (TPA) to predict the PJK/F. External validation is required to verify its clinical usefulness. Materials and Methods. We included patients with ASD undergoing ≥ 5-level fusion including the sacrum or pelvis. SAAS was calculated based on the scores of the three components: PI-LL, PT, and TPA. PJK/F rates and clinical outcomes were compared among the correction categories (undercorrection, matched correction, and overcorrection) for the SAAS as well as for each of the three components. PJK/F rates were compared according to the correction groups of the sagittal components and total SAAS using the χ² test. Receiver operating characteristic (ROC) analysis was performed to evaluate the predictive ability of overcorrection to develop PJK/F for the three sagittal parameters and SAAS. PROMs at final follow-up were compared among correction groups using ANOVA with Bonferroni post hoc corrections. Results. A total of 411 patients were included in the study (mean age: 69.3 y, mean body mass index: 25.9 kg/m², total levels fused: 7.7 levels, and follow-up duration: 43.3 mo). Postoperative SAAS categories were as follow: undercorrection (13.4%), matched correction (30.2%), and overcorrection (56.4%). The PJK/F rates were significantly higher in the overcorrection group relative to PI-LL component (P = 0.001) as well as SAAS (P = 0.038) compared with undercorrection or matched correction groups. The clinical outcomes were best in patients who achieved matched correction relative to PI-LL component as well as SAAS compared with the other correction groups. However, the differentiating power of clinical outcomes across the correction categories was greater in the PI-LL component than in the SAAS. Conclusion. This study validated the efficacy of SAAS system to differentiate PJK/F development and good clinical outcomes. However, its differentiating power seems to be largely attributable to the function of the PI-LL component, as the PI-LL correction status better predicted PJK/F risk and clinical outcomes than SAAS.