Comparative performance of the nutritional risk index and modified 5-item frailty index in predicting surgical outcomes: a multi-institutional analysis of 9.7 million patients
Background and Aims - Preoperative risk stratification is crucial for surgical decision-making and patient counseling. While the Modified 5-Item Frailty Index (mFI-5) is widely used, the Nutritional Risk Index (NRI) has emerged as a promising alternative due to its dynamic and continuous nature. How...
Saved in:
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Article (Journal) |
| Language: | English |
| Published: |
January 2026
|
| In: |
Clinical nutrition
Year: 2026, Volume: 56, Pages: 1-9 |
| ISSN: | 1532-1983 |
| DOI: | 10.1016/j.clnu.2025.11.020 |
| Online Access: | Verlag, lizenzpflichtig, Volltext: https://doi.org/10.1016/j.clnu.2025.11.020 Verlag, lizenzpflichtig, Volltext: https://www.sciencedirect.com/science/article/pii/S0261561425003152 
|
| Author Notes: | Samuel Knoedler, Sarah Friedrich, Kevin Hu, Leonard Knoedler, Thomas Schaschinger, Felix Klimitz, Jasper Iske, Gabriel Hundeshagen, Adriana C. Panayi, Bong-Sung Kim, Martin Kauke-Navarro, Fortunay Diatta, Bohdan Pomahac |
| Summary: | Background and Aims - Preoperative risk stratification is crucial for surgical decision-making and patient counseling. While the Modified 5-Item Frailty Index (mFI-5) is widely used, the Nutritional Risk Index (NRI) has emerged as a promising alternative due to its dynamic and continuous nature. However, their comparative effectiveness remains unclear. - Objective - To compare the predictive performance of the NRI versus the mFI-5 for predicting postoperative complications, readmissions, and mortality in a large, multi-institutional surgical cohort. - Methods - We analyzed 9,782,974 patients from the American College of Surgeons National Surgical Quality Improvement Program database. Three predictive models using NRI (dichotomous, spline, and adjusted) and two using mFI-5 (unadjusted and adjusted) were developed for six outcomes: overall complications, surgical complications, medical complications, readmission, unplanned reoperation, and mortality. Model discrimination was assessed using the area under the receiver operating characteristic curve (AUC), and calibration was evaluated using bootstrap-validated calibration plots. - Results - The overall complication rate was 14.7 % (n = 1,442,984). Both adjusted NRI and mFI-5 models showed excellent discrimination for mortality (AUC = 92.6 for both) and medical complications (AUC = 83.7 and 83.5, respectively). The NRI also demonstrated comparable or superior performance for surgical complications (AUC = 75.8 vs 75.4), readmission (AUC = 69.8 vs 71.3), and unplanned reoperation (AUC = 65.2 vs 65.0). Mortality discrimination was 92.6 for both models. Notably, NRI values were significantly lower in patients who experienced complications compared to those who did not (96.0 vs 103, p < 0.0001). - Conclusions - The NRI demonstrates comparable, and in some cases, superior, predictive performance to mFI-5 across surgical outcomes. As a dynamic and continuous measure, NRI may offer advantages over mFI-5 for preoperative risk stratification. These findings support the integration of NRI into surgical risk assessment protocols. |
|---|---|
| Item Description: | Online verfügbar: 17. Dezember 2025, Artikelversion: 26. Dezember 2025 Gesehen am 13.03.2026 |
| Physical Description: | Online Resource |
| ISSN: | 1532-1983 |
| DOI: | 10.1016/j.clnu.2025.11.020 |