Frailty as a Predictor of Failure to Rescue in Spine Surgery: A Retrospective Analysis of a National Surgical Database

Introduction: Frailty can impact spine surgery outcomes. This study aims to assess the abilities of the Risk Analysis Index (RAI) to predict failure to rescue (FTR) in spine surgery patients.

Methods: An observational, retrospective study of the ACS-NSQIP database from 2015 – 2020 yielded 350,469 spinal surgery patients. Outcomes including FTR, non-routine discharge (NRD), and extended length of stay (eLOS) were examined in patients with major complications. Area Under the Receiver Operating Characteristic Curve (AUC)/C-statistic was measured for RAI alone, RAI combined with the American Society of Anesthesiologists (ASA) classification, and later, ASA and number of postoperative complications.

Results: Major complications occurred in 25,471 spine surgery patients and 974 (3.8%) experienced death within 30 days of surgery, or failure to rescue (FTR). Patient distribution across increasing frailty categories (RAI I-IV) was: I – 5.5%, II – 36.7%, III – 45.3%, and IV – 12.5%. For patients with increasing frailty, the adjusted odds ratio (aOR) for mortality increased from 1.604 for RAI II to 11.317 for RAI IV. Similar trends were observed when analyzing NRD and eLOS. AUC were 0.740 (0.723-0.756) for RAI, 0.757 (0.741-0.772) for RAI and ASA, and 0.861 (0.851-0.872) for RAI combined with ASA and the number of mortality complications, 0.658 (0.651-0.665), 0.671 (0.665-0.678), and 0.696 (0.690-0.703) for NRD, and 0.581 (0.572-0.589), 0.632 (0.624-0.640), 0.691 (0.683-0.699) for eLOS, respectively for RAI, RAI with ASA, and lastly combination of RAI, ASA, and number of complications experienced.

Conclusion : Frailty demonstrated strong discrimination in predicting FTR after major complications in spine surgery patients. The combination of RAI, ASA, and complication count provided superior discrimination to RAI or RAI-ASA alone for identifying patients at high-risk for experiencing FTR. These findings can enable better identification of high-risk patients to provide more precise risk/benefit discussions of potential spine surgery in order to reduce mortality and adverse outcomes.