PP151 - PROGNOSTIC BURDEN OF GLIM-DEFINED MALNUTRITION IN THE ICU: FIRST NATIONWIDE CLAIMS-DATABASE EVIDENCE
PP151
PROGNOSTIC BURDEN OF GLIM-DEFINED MALNUTRITION IN THE ICU: FIRST NATIONWIDE CLAIMS-DATABASE EVIDENCE
Y. Ishida1,2,*, K. Maeda1,2,3,4, T. Inoue5, K. Murotani6,7, A. Saito5,8, F. Kawase9, A. Nagano10, N. Mori1,3,11
1Department of Nutrition, Aichi Medical University Hospital, 2Department of Palliative and Supportive Medicine, Graduate School of Medicine, Aichi Medical University, 3Nutrition Therapy Support Center, Aichi Medical University Hospital, 4Department of Geriatric Medicine, Hospital, National Center for Geriatrics and Gerontology, Aichi, 5Graduate School of Health and Welfare, Niigata University of Health and Welfare, Niigata, 6School of Medical Technology, 7 Biostatistics Center, Kurume University, Fukuoka, 8Department of Nutrition, Yoshida Hospital, Niigata, 9Department of Sports and Health Sciences, Graduate School of Biomedical Engineering, Tohoku University, Miyagi, 10Caresso, Hyogo, 11Department of Palliative and Supportive Medicine, Aichi Medical University, Aichi, Japan
Rationale: The Global Leadership Initiative on Malnutrition (GLIM) criteria were established to standardise malnutrition diagnosis worldwide, yet intensive care unit (ICU) evidence rests entirely on single-centre or point-prevalence survey data, and no study has validated GLIM severity grading for ICU outcomes using nationwide real-world data.
Methods: Using a Japanese nationwide claims database covering 6 years and 7 months from April 2018, adults aged 18 to 99 years at first ICU admission with available 6-month bodyweight records were enrolled in a retrospective cohort. GLIM phenotypic criteria—weight loss, body mass index, and estimated appendicular skeletal muscle mass index—assigned three nutritional severity tiers. Fine–Gray (subdistribution hazard ratio [sHR]) and Cox (hazard ratio [HR]) regression addressed in-hospital and overall mortality; hospital and ICU lengths of stay by restricted mean survival time, adjusted for age, sex, Sequential Organ Failure Assessment score, surgery, and comorbidity.
Results: Of 7,460 assessed, 6,065 were included (age 74.4±12.7 years; 36.4% female). GLIM criteria classified 85.4% as malnourished. In-hospital mortality rose from 7.5% (no malnutrition) to 13.4% (severe; sHR 1.58 [1.16–2.15]; p=0.004). All-cause mortality followed a dose-response gradient: 14.6%, 21.4% (moderate; HR 1.28 [1.06–1.56]; p=0.013), and 25.1% (severe; HR 1.62 [1.32–1.99]; p<0.001). Hospital stay was extended by 1.55 days (moderate; p=0.011) and 4.09 days (severe [2.72–5.46]; p<0.001); ICU stay was unaffected.
Conclusion: In this first nationwide claims-database validation of GLIM criteria among ICU patients, each severity tier independently increased mortality and prolonged hospitalisation in a dose-response manner, providing real-world evidence to support integrating GLIM-based nutritional assessment into routine ICU admission practice.
Disclosure of Interest: None declared