Comparison of the Geriatric Nutritional Risk Index and the Risk Analysis Index for Prediction of Operation Delay in Adult Spinal Deformity Surgery

Introduction: Increasing spine research investigates predictive indices for adverse postoperative outcomes in adult spinal deformity (ASD) surgery, which places a significant physiologic burden on increasingly older patients. This study sought to compare the predictive ability of the Risk Analysis Index (RAI) with the Geriatric Nutritional Risk Index (GNRI) for operation delay in ASD patients.

Methods: The ACS National Surgical Quality Improvement Program (NSQIP) database was queried for ASD surgery patients using diagnostic and procedural codes (2015-2020). RAI is a 9-factor, 14-item prospective assessment of frailty and functional status. GNRI uses preoperative albumin and a patient’s weight relative to their ideal body weight to quantify nutritional status. Primary outcomes included operation delay, defined as the top 95% of days from admission to operation, and mortality. Secondary outcomes included non-home discharge (NHD), extended length of stay (eLOS), and Clavien-Dindo Grade IV (CDIV) complications. The comparative discriminatory accuracy of GNRI and RAI for operation delay was determined using multivariate modeling and area under the receiver operating characteristic curve (AUROC) analysis.

Results: A total of 2,369 patients undergoing ASD surgery were identified (58.3% female, median age=64 years, IQR:54-71). Adverse postoperative outcomes included NHD (38.4%, N=920), eLOS (24.7%, N=593), CDIV complications (8.5%, N=204), and mortality (1.0%, N=24). Operation delay (>3days) was recorded in 141 (5.9%) patients. GNRI demonstrated robust discriminatory accuracy for prediction of operation delay in AUROC analysis (C-statistic: 0.73, 95% CI:0.68−0.76, p< 0.001) compared to RAI (C-statistic: 0.59, 95% CI:0.55−0.63, p< 0.001) and mFI-5 (C-statistic: 0.57, 95% CI:0.52−0.62, p< 0.001)(DeLong test, p< 0.01). In multivariable logistic regression, adjusted for potential confounders including patient demographics, RAI, GNRI, ASA, and PASC, among others, patients who experienced operation delay had increased risk for mortality (OR: 3.38, 95% CI:1.08-10.63, p=0.037), NHD (OR=2.53, 95% CI:1.66-3.86, p< 0.001), eLOS (OR=2.80, 95% CI:1.55-5.05, p< 0.001), and CDIV complications (OR=2.44, 95% CI:1.50-3.97, p< 0.001).

Conclusion : GNRI demonstrates superior discriminatory accuracy than RAI for prediction of operation delay in ASD surgery, which is associated with greatest increased risk for mortality and eLOS. Patients admitted to the hospital for ASD surgery may undergo preoperative risk stratification using GNRI to inform perioperative management and promote timely surgical intervention.