Abstract
Study DesignRetrospective cohort study.ObjectiveTo evaluate the predictive performance (discrimination and calibration) of the mFI-5 for major postoperative complications in thoracolumbar spine surgery, and to compare its effectiveness in low-, moderate-, and high-complexity procedures.MethodsThe study was conducted on adult patients (>18 years) who underwent thoracolumbar spine surgery at a single tertiary care center between 2017 and 2020. The primary outcome was the incidence of major complications within 90 days. Model discrimination was assessed using the area under the receiver operating characteristic curve (AUROC), and calibration was evaluated with calibration-in-the-large (CITL), calibration slope, and the Hosmer-Lemeshow test.ResultsA total of 839 patients were included (mean age: 62.8 years; SD: 16.8). Major complications occurred in 8.2% of cases. The mFI-5 demonstrated fair discrimination (AUROC: 0.66; 95% CI: 0.60-0.70) and excellent calibration (slope = 1, CITL = 0; Hosmer-Lemeshow P = .99). Stratified analysis showed improved discrimination in high-complexity surgeries (AUROC: 0.74; 95% CI: 0.64-0.84), compared to moderate (0.62; 95% CI: 0.48-0.74) and low complexity (0.63; 95% CI: 0.50-0.74) procedures. Readmission rates were 7% at 30 days and 9% at 90 days, with a 6-month mortality rate of 1%.ConclusionThe mFI-5 is a valuable tool for predicting major complications in thoracolumbar spine surgery, particularly in high-complexity procedures. Its predictive performance is limited in lower-complexity surgeries. Further prospective studies are needed to validate its use and enhance preoperative risk stratification.