2026 Proffered Presentations
S271: THE RISK ANALYSIS INDEX OUTPERFORMS THE MODIFIED FRAILTY INDEX IN PREDICTING POSTOPERATIVE COMPLICATIONS IN SKULL BASE SURGERY
Akshay K Warrier1; Ryan Bartholomew, MD2; Daniel J Lee, MD, FACS2; 1Rutgers New Jersey Medical School; 2Massachusetts Eye and Ear
Objective: To compare the performance of the Risk Analysis Index (RAI) and 5-item Modified Frailty Index (mFI-5) in predicting 30-day postoperative outcomes following skull base surgery.
Study Design: Retrospective cohort study.
Methods: Adult patients undergoing skull base procedures were identified in the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) from 2005 to 2020. Frailty was assessed using both RAI and mFI-5 (Figure 1). Patients were stratified into non-frail, pre-frail, frail, and severely frail categories based on index-specific thresholds. Primary outcomes included 30-day mortality, Clavien–Dindo class II (CD II) and class IV (CD IV) complications, extended length of stay (eLOS), non-home discharge (NHD), and deep surgical site infection (DSSI). Multivariable logistic regression models assessed the independent association of frailty with outcomes. Predictive performance was quantified using area under the receiver operating characteristic curve (AUC).
Results: A total of 2,809 skull base surgery patients were included (mean age 54.1 years, range 18–89). RAI-defined frailty was significantly associated with higher odds of CD II (OR 10.24, 95% CI 3.58–29.24), CD IV (OR 3.56, 95% CI 1.36–9.31), eLOS (OR 6.34, 95% CI 3.40–11.83), NHD (OR 18.20, 95% CI 6.35–52.17), and DSSI (OR 9.22, 95% CI 2.40–35.45) (all p<0.05). RAI also predicted 30-day mortality (OR 14.47, 95% CI 3.02–69.31). In contrast, mFI-5 was predictive of CD II (OR 2.79, 95% CI 1.88–4.16), eLOS (OR 3.55, 95% CI 1.43–8.82), and NHD (OR 3.80, 95% CI 1.73–8.37), but was not significantly associated with CD IV, DSSI, or mortality.
ROC curve analysis confirmed similar discrimination across nearly all outcomes. RAI achieved slightly higher AUCs than mFI-5 for mortality (0.722 vs. 0.690), CD II (0.793 vs. 0.788), CD IV (0.704 vs. 0.700), eLOS (0.692 vs. 0.696), and NHD (0.721 vs. 0.720). AUCs were similar for CD I (0.723 vs. 0.728), DSSI (0.649 vs. 0.627), OSSI (0.648 vs. 0.646), and wound dehiscence (0.758 vs. 0.771).
Conclusion: RAI outperforms mFI-5 in predicting critical postoperative outcomes following skull base surgery, including mortality, serious complications, and discharge disposition, while offering similar discriminative ability. Unlike the comorbidity-based mFI-5, RAI integrates functional status, nutritional decline, and physiologic reserve, offering a more comprehensive assessment of patient vulnerability. These findings support the integration of RAI into preoperative workflows for skull base procedures and suggest its potential as a standard risk stratification tool in skull base and complex cranial surgery. Prospective validation is warranted to assess clinical impact on surgical decision-making, perioperative planning, and patient-centered outcomes.
