Validation of modified GLIM criteria to predict adverse clinical outcome and response to nutritional treatment: A secondary analysis of a randomized clinical trial

Background & aims The Global Leadership Initiative on Malnutrition (GLIM) recently suggested specific criteria to standardize the diagnosis of malnutrition. There is need for validation of these criteria regarding response to nutrition treatment. Our aim was to validate modified GLIM (mGLIM) criteria among medical inpatients at risk of disease related malnutrition for prediction of outcome and response to nutritional therapy. Methods This is a secondary analysis of the Effect of Early Nutritional Support on Frailty, Functional Outcomes, and Recovery of Malnourished Medical Inpatients Trial (EFFORT), a multicenter randomized controlled trial conducted between April 2014 and February 2018. Adult medical inpatients at nutritional risk (Nutrition Risk Score 2002 ≥ 3 points) were randomly assigned to receive nutritional therapy according to an algorithm based on individualized nutritional requirements (intervention group) or standard hospital food (control group). We included all participants with available information regarding mGLIM criteria. The primary outcome was adverse clinical outcome, which was a composite of 30-day all-cause mortality, ICU-admission, rehospitalization rate, major complications and decline in functional status. Results Of 1917 eligible participants at nutritional risk, 1181 (61.6%) met the diagnosis of malnutrition based on mGLIM criteria. The incidence of adverse clinical outcome was significantly higher in mGLIM-positive participants compared with mGLIM-negative participants [330/1181 (27.9%) versus 140/736 (19.0%); multivariable adjusted odds ratio [OR] 1.53; 95% CI 1.22–1.93; p < 0.001]. Regarding the effect of nutritional therapy, the reduction in adverse clinical outcomes was higher in mGLIM-positive participants [180/581 (31.0%) vs. 150/600 (25.0%), OR 0.69; 95% CI 0.53–0.9, p = 0.007], compared with mGLIM-negative participants [75/379 (19.8%) versus 65/357 (18.2%), OR 0.95; 95% CI 0.65–1.40, p = 0.797], a finding that was, however, not significant in interaction analysis (p for interaction = 0.217). Conclusion Data from this secondary analysis of a multicenter randomized trial involving medical inpatients at nutritional risk validate the strong prognostic value of mGLIM criteria regarding adverse clinical outcomes and other long-term outcomes. However, further research is needed to improve the ability of GLIM criteria to predict therapeutic response to nutritional interventions