Inconsistent Associations of Modified Frailty Index-5 With Adverse Head and Neck Reconstruction Outcomes

Abstract

Objectives: To investigate the consistency of associations between modified frailty index-5 (mFI-5) and postoperative adverse outcomes in head and neck cancer (HNC) reconstruction. Methods: American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database from 2017 to 2022 was utilized to identify HNC patients undergoing locoregional or microvascular free tissue transfers. Kaplan–Meier estimates and multivariable Cox regression analyses were utilized to compare risk of infections, bleeding, readmissions, reoperations, major adverse cardiovascular events (MACE), and mortality within the first postoperative month for each mFI-5 score with mFI-5 = 0 as reference. Further analyses investigated associations between individual mFI-5 components and the outcomes of interest. Results: We included 5,573 patients (median age: 64; 31.5% female), 63% (n = 3,519) of whom underwent microvascular free tissue transfers. Unadjusted univariate analyses associated higher mFI-5 scores with longer hospital stays. In locoregional tissue transfers, adjusted hazard ratios (aHRs) for reoperation were 1.37 (p = 0.03) for mFI-5 = 1 and 2.19 (p = 0.03) for mFI-5 ≥ 3. In microvascular free tissue transfers, aHRs for MACE were 1.93 (p = 0.04) for mFI-5 = 2 and 6.53 (p < 0.001) for mFI-5 ≥ 3, while aHRs for mortality was 3.88 (p = 0.04) for mFI-5 ≥ 3. No associations were observed between increasing mFI-5 scores and increased relative risk of infection, bleeding, or readmission. Individual component analysis associated congestive heart failure with aHRs of 3.92 (1.84–8.35; p < 0.001) for MACE and 5.30 (2.03–13.88; p < 0.001) for mortality. Additionally, COPD was associated with an aHR of 1.39 (1.16–1.67; p < 0.001) for infections. Conclusion: The associations of higher mFI-5 scores with postoperative adverse outcomes following oncoplastic head and neck reconstruction were inconsistent and possibly driven by individual effects of its components. Level of Evidence: 3 Laryngoscope, 135:2342–2352, 2025.