Predictive performance of GLIM with different nutritional risk screening tools in 5-year adverse health outcomes among older hospitalized patients: a comparative study

Abstract

Background: Global Leadership Initiative on Malnutrition (GLIM) has proposed a two-step approach for malnutrition diagnosis: screening with any validated nutritional screening tool (NST) to identify ‘at-risk’ patients, followed by detailed nutritional assessment to diagnose malnutrition severity and grade it. Given the existence of multiple validated NSTs, this study aims to compare the predictive relationships between GLIM and different nutritional screening tools on 5-year adverse health outcomes (mortality, readmission, and falls) among elderly hospitalized patients. Methods: Retrospective, single-center, observational study from January 2015 to February 2020.A total 2001 elderly hospitalized patients. Multiple nutritional screening tools were evaluated at admission. Nutritional status was assessed using GLIM criteria.Cox regression and logistic regression analysis was performed to examine the predictive performance of GLIM criteria on 5-year adverse health outcomes (mortality, readmission, and falls). Results: The prevalence of malnutrition varied from 16.44% to 46.08%, depending on the nutritional screening tool used for classification. During the 5-year follow-up period, 203 patients (10.14%) died, 458 patients (22.88%) experienced at least one fall, and 995 patients (49.73%) were readmitted to hospital at least once. GLIM criteria effectively predicted 5-year mortality and readmission rates, with both GLIM and MNA-SF screening demonstrating the best predictive performance [Hazard ratios: 3.348 (2.781–4.029),Odds ratios 3.933 (2.953–5.239)]. Conclusion: Malnutrition identified by both GLIM and MNA-SF demonstrates robust predictive ability for mortality and readmission rates in elderly hospitalized patients. The GLIM criteria’s predictive capacity for mortality and readmission may guide interventions, holding significant implications for both clinical practice and research.