-
Modified Global Leadership Initiative on Malnutrition definitions in mechanically ventilated patients: diagnostic performance and association with 90-day mortality
-
Ah Ron Lee
, Bo-Eun Kim , Eun-Mee Kim , Chi-Min Park , Sung Nim Han
-
Received May 23, 2026 Accepted August 13, 2026 Published online November 11, 2026
-
DOI: https://doi.org/10.15747/ACNM.26.0058
-
-
Abstract
- Purpose
The Global Leadership Initiative on Malnutrition (GLIM) criteria can be difficult to apply during initial assessment of critically ill patients. This study evaluated the agreement and diagnostic performance of clinically feasible, modified GLIM approaches and their associations with mortality.
Methods This retrospective cohort study included critically ill adults (≥18 years) admitted to the intensive care unit who received invasive mechanical ventilation. Nutritional status was assessed using the Subjective Global Assessment, the Academy of Nutrition and Dietetics/American Society for Parenteral and Enteral Nutrition criteria, and two modified GLIM-based definitions: GLIM1 omitted muscle mass assessment, whereas GLIM2 used the Clinical Frailty Scale (CFS) as a proxy. Agreement and accuracy measures were used to compare GLIM1 and GLIM2 with reference classifications; regression models assessed associations with mortality.
Results The study included 595 patients (mean age, 63.2±14.3 years; 64% male). Malnutrition prevalence was 43% with GLIM1 and 64% with GLIM2 (n=457 with available CFS). GLIM1 had high specificity (83%–84%) but moderate sensitivity (61%–64%). GLIM2 had greater sensitivity (84%–85%) and agreement (κ=0.52–0.59), but lower specificity (68%–75%). In fully adjusted models, GLIM1 remained associated with 90-day mortality (hazard ratio [HR], 1.45; 95% confidence interval [CI], 1.15–1.81), whereas the GLIM2 association was attenuated (HR, 1.24; 95% CI, 0.94–1.63).
Conclusion The simplified, modified GLIM approaches were clinically feasible and showed moderate agreement, although diagnostic performance was modest. The more consistent association of GLIM1 than GLIM2 with mortality requires further confirmation.
|