Challenges and limitations of the existing scoring systems for postoperative outcomes in thoracolumbar pyogenic spondylitis: a multicenter retrospective cohort study.

Abstract

Purpose: This study aimed to evaluate the predictive ability of existing scoring systems for postoperative outcomes in patients with thoracolumbar pyogenic spondylitis treated with minimally invasive posterior fixation.

Methods: We conducted a multicenter retrospective cohort study of 90 patients with thoracolumbar pyogenic spondylitis treated with minimally invasive posterior fixation between January 2014 and June 2024. We assessed the Brighton spondylodiscitis score (BSDS), spinal instability spondylodiscitis score (SISS), and spinal infection treatment evaluation (SITE) score. Patients were divided into success and failure groups based on infection control or implant failure. The predictive ability of each scoring system for treatment failure was evaluated using receiver operating characteristic (ROC) curves and area under the curve (AUC).

Results: The success and failure groups consisted of 78 (87%) and 12 patients (13%), respectively, with the failure group including seven and five patients due to infection control or implant failure. The failure group had higher BSDS, SISS, and SITE scores, but the ROC analysis showed low predictive accuracy (AUC: BSDS = 0.65, SISS = 0.64, SITE score = 0.56). Grading the BSDS into three categories revealed no failure in the low-risk patients, whereas 22% of the high-risk patients required unplanned additional surgeries for infection control or implant failure.

Conclusions: Existing scoring systems showed limited ability to predict postoperative outcomes in patients with thoracolumbar pyogenic spondylitis treated with minimally invasive posterior fixation. BSDS grading demonstrated some utility, with 22% of high-risk patients experiencing postoperative infection control failure and implant failure, highlighting the need for careful treatment planning.