Optimization of Frailty Indices in Predicting the Risk of Surgical Outcomes in Patients Undergoing Spine Surgery Under Non-General Anesthesia.

Abstract

Non-general anesthesia (NGA) techniques are increasingly utilized in spine surgery for potential perioperative benefits. However, reliable approaches to risk stratification in this population remain undefined. Frailty indices are widely used across surgical specialties, yet their predictive accuracy and optimal integration within NGA spine cohorts have not been fully established. To evaluate the predictive performance of common frailty indices in adults undergoing spine surgery under NGA, and to determine whether combining frailty measures with perioperative clinical variables (particularly American Society of Anesthesiologists classification [ASA]) improves prediction of short-term postoperative outcomes. Retrospective cohort study using the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database (2010-2022). Adult patients undergoing spine surgery under NGA identified in ACS-NSQIP (2010-2022). Thirty-day postoperative (1) complications, (2) readmission, and (3) reoperation. Frailty was quantified using the 5-Item Modified Frailty Index (mFI-5), 11-Item Modified Frailty Index (mFI-11), and Risk Analysis Index-Administrative (RAI-A). Each index was evaluated independently and in combination with ASA classification and body mass index (BMI). Multivariable logistic regression models were constructed for each outcome. Model performance was assessed using area under the receiver operating characteristic curve (AUC), integrated discrimination improvement (IDI), and net reclassification improvement (NRI), with 1,000 bootstrap resamples. Among 3,019 patients undergoing spine surgery under NGA, 30-day complication, readmission, and reoperation rates were 6.0%, 3.4%, and 1.7%, respectively. The combined mFI-11 + ASA model demonstrated superior discrimination for complications (AUC 0.751 vs 0.578; ΔAUC = 0.172), readmission (AUC 0.660 vs 0.474; ΔAUC = 0.186), and reoperation (AUC 0.618 vs 0.541; ΔAUC = 0.077), with significant IDI and NRI improvement. In multivariable analysis, both mFI-11 and ASA independently predicted complications and readmission, whereas ASA alone was associated with reoperation in a dose-dependent manner. Perioperative status as reflected by ASA classification is critical for risk assessment in patients undergoing spine surgery under NGA. Integrating ASA into mFI-11 significantly enhances short-term prediction. Higher mFI-11 frailty scores and ASA classes were associated with increased risks of complications and readmissions, underscoring the importance of careful patient selection and preoperative optimization when considering NGA in frail individuals. The proposed risk equation offers a practical approach for risk quantification and supports individualized surgical planning.