Background Frailty indices are well-established metrics used to assess physiologic reserve and predict peri- and post-operative morbidity and mortality. While multiple frailty tools exist, there is limited consensus regarding the most effective risk assessment tool in orthopaedic trauma. Although there is literature supporting the utility of the 5-Item Modified Frailty Index (mFI-5) in the orthopaedic trauma population, the newer Revised Risk Analysis Index (RAI-Rev) has not been evaluated in this population. This study sought to determine the applicability of RAI-Rev in the orthopaedic trauma population and compare the predictive discrimination for the RAI-Rev and mFI-5 for 30-day postoperative outcomes.

Methods The ACSNSQIP data were queried from 2015 to 2020 for patients ≥ 18 with operatively treated fractures of the forearm, humerus, pelvis, acetabulum, femur, tibia, and hindfoot. Outcomes included a composite of major complications (death, sepsis, pulmonary embolism, ventilator use greater than 48 hours, return to operating room, unplanned re-intubation, acute postoperative renal failure, cardiac arrest, myocardial infarction, cerebrovascular accident, and graft failure), mortality, readmission, and wound complications. Frailty scores were stratified into four cohorts based on frailty levels, utilizing previously described cutoff scores. Multivariate analysis was used to assess the predictive value of frailty. Receiver operating characteristics curve analysis was quantified with C-statistic for discriminatory accuracy.

Results In total, 206,352 patients met inclusion criteria. Median age was 69 years old, with 64.2% (n = 132,514) being female. Multivariate regression analysis showed that increasing frailty tiers in both RAI-Rev and mFI-5 were independent predictors of mortality, major complications, readmission, and wound complications. The cohort with the highest degree of frailty in both RAI-Rev and mFI-5 had the greatest risk of poor outcomes. RAI-Rev had significantly superior predictive discriminatory thresholds compared to mFI-5 for predicting 30-day mortality (C-statistic: RAI-Rev (0.84) and mFI-5 (0.67), p <0.001), major complications (C-statistic: RAI-Rev (0.73) and mFI-5 (0.65), p <0.001), and readmission (C-statistic: RAI-Rev (0.68) and mFI-5 (0.63), p <0.001). However, mFI-5 outperformed RAI-Rev when predicting wound complications (C-statistic: RAI-Rev (0.52) and mFI-5 (0.55), p <0.001).

Conclusion RAI-rev demonstrated superior predictability of postoperative morbidity, mortality, and readmission rates compared to mFI-5, but was less effective in predicting surgical site complications. These findings demonstrate the utility of RAI-Rev in anticipating postoperative complications in the setting of orthopaedic trauma, where optimizing surgical candidate selection is not always possible. Assessing the predicted morbidity and mortality through RAI-Rev may allow surgeons to mitigate postoperative complications in patients undergoing surgical fixation of traumatic injuries.