Comparative Analysis of ACS Surgical Risk Calculator, Modified Frailty Index, Geriatric Nutritional Risk Index, Risk Analysis Index, and Albumin for Predicting Complications in Elective Spine Surgery

Introduction: Frailty and nutritional status impact outcomes in spine surgery. Indices like the Modified Frailty Index-5 (mFI-5), Geriatric Nutritional Risk Index (GNRI), Risk Analysis Index (RAI), preoperative albumin (PRALBUM), and the ACS Surgical Risk Calculator are used to predict complications. This study compares these indices' predictive abilities in elective spine surgery.

Methods: A retrospective analysis using the 2019–2020 ACS NSQIP database identified 55,495 patients undergoing elective spine surgery. mFI-5, GNRI, RAI, PRALBUM, and ACS Surgical Risk Calculator scores were calculated. Receiver Operating Characteristic (ROC) curves for 21 complications were generated to assess sensitivity, specificity, AUC, and optimal thresholds for each index.

Results: The ACS Surgical Risk Calculator showed the highest predictive ability for severe complications. For end-of-life care withdrawal, it achieved an AUC of 0.946 (sensitivity 88.5%, specificity 87.5%). It also demonstrated high predictive value for prolonged hospitalization (AUC 0.890, sensitivity 80.0%, specificity 81.4%), ventilator weaning failure (AUC 0.883, sensitivity 79.9%, specificity 79.6%), septic shock (AUC 0.874, sensitivity 80.7%, specificity 80.3%), unplanned intubation (AUC 0.839, sensitivity 74.8%, specificity 75.6%), renal insufficiency (AUC 0.837, sensitivity 74.6%, specificity 75.7%), pneumonia (AUC 0.824, sensitivity 72.0%, specificity 76.6%), and cerebrovascular accident (AUC 0.789, sensitivity 76.8%, specificity 68.3%).

PRALBUM was the best predictor for immediate complications, including reoperation due to renal failure (AUC 0.773, sensitivity 80.6%, specificity 67.7%), cardiac arrest (AUC 0.731, sensitivity 68.8%, specificity 66.7%), and deep vein thrombosis (AUC 0.621, sensitivity 61.4%, specificity 57.9%). RAI effectively predicted frailty-related systemic complications such as cerebrovascular accident (AUC 0.753, sensitivity 69.7%, specificity 67.9%), septic shock (AUC 0.726, sensitivity 61.7%, specificity 70.4%), and pneumonia (AUC 0.707, sensitivity 64.8%, specificity 65.6%). GNRI and mFI-5 showed moderate predictive ability (AUC 0.221-0.691 and 0.328-0.644, respectively).

Conclusion : The ACS Surgical Risk Calculator effectively predicts mortality-related outcomes and severe complications in elective spine surgery. PRALBUM is valuable for immediate postoperative risks, while RAI predicts frailty-related complications. Combining ACS Surgical Risk Calculator, PRALBUM, and RAI can improve risk stratification and clinical decision-making.