Assessing Frailty Screening Tools for Predicting Morbidity and Mortality in Spine Surgery for Spinal Epidural Abscess

Introduction: The revised risk analysis index (RAI-rev) and 5-factor modified frailty index (mFI-5) are comprehensive assessment tools of frailty that have been used to predict neurosurgical outcomes. The aim of this study is to investigate the utility of these tools to predict extended hospital length of stay (LOS), non-routine discharge (NRD), 30-day hospital readmission, and mortality among patients undergoing surgery for spinal epidural abscess (SEA).

Methods: We conducted a retrospective cohort study using the 2011-2022 ACS National Surgical Quality Improvement Program (NSQIP) database to identify adult patients who underwent spinal surgery for SEA. Using receiver operating characteristic and multivariable analyses, we compared the discriminative thresholds and independent associations of RAI-rev, mFI-5, and greater patient age with extended LOS, NRD, 30-day hospital readmission, and mortality.

Results: Among the 1,914 patients studied, 734 (38.4%) were female, and 1,198 (72.4%) were non-Hispanic White. According to RAI-rev, 1,043 (54.5%) patients were categorized as robust, 692 (36.2%) as typical, and 179 (9.4%) as frail. In contrast, mFI-5 categorized 745 (38.9%) as robust patients, 640 (33.4%) as typical, and 529 (27.6%) as frail. In predicting extended LOS, RAI-rev (AUC 0.5509) outperformed patient age (p < 0.001) but was comparable to mFI-5. Similarly, RAI-rev (AUC 0.6859) demonstrated superior prediction of 30-day mortality compared to patient age (p = 0.033) but was comparable to mFI-5. With respect to 30-day readmission, mFI-5 (AUC 0.5706) outperformed RAI-rev (p < 0.001). No differences were observed between indices for predicting NRD. On multivariable analysis, RAI-rev was a significant predictor of extended LOS [aOR: 1.60, 95% CI: 1.01–2.54] and NRD [aOR: 1.97, 95% CI: 1.23–3.15] for frail patients. However, RAI-rev was not significantly associated with 30-day readmission or mortality. For frail patients, mFI-5 was a significant predictor for NRD [aOR: 1.66, 95% CI: 1.26–2.19] and 30-day readmission [aOR: 2.11, 95% CI: 1.48–3.02] but not significantly associated with extended LOS nor 30-day mortality.

Conclusion : Our study shows that RAI-rev and mFI-5 moderately predict morbidity and mortality in patients undergoing surgery for spinal epidural abscess. More robust tools are needed for better risk stratification to improve patient care and resource utilization.