PD329 - NUTRITIONAL RISK SCREENING WITH NRS-2002 IN GASTROENTEROLOGY: PREDICTORS OF NUTRITIONAL RISK, CONCORDANCE WITH GLIM CRITERIA, AND IMPACT ON LENGTH OF HOSPITAL STAY

Linked sessions

PD329

NUTRITIONAL RISK SCREENING WITH NRS-2002 IN GASTROENTEROLOGY: PREDICTORS OF NUTRITIONAL RISK, CONCORDANCE WITH GLIM CRITERIA, AND IMPACT ON LENGTH OF HOSPITAL STAY

L. Pozzi1,*, D. Giustozzi2, L. Elli1, F. Caprioli1

1Ospedale Maggiore Policlinico di Milano, 2Università degli Studi di Milano , Milano , Italy

 

Rationale: Hospital malnutrition is a clinically and economically significant challenge. The NRS-2002 is the ESPEN gold standard for nutritional screening in hospitalised patients; since January 2024, the Lombardy Region has mandated its systematic use in all hospitals. The aim was to assess the prevalence of nutritional risk, identify independent predictors, and quantify the impact on length of hospital stay (LOS) in a Gastroenterology unit.

Methods: Retrospective observational study conducted in the Gastro-Hepatology and Gastro-Endoscopy wards of Policlinico di Milano (June 2023–2024). 290 consecutive patients were included. Variables collected: NRS-2002, anthropometric parameters, admission laboratory values, diagnoses, LOS. Some of the GLIM criteria were recalculated from raw data. Statistical analyses: Kruskal-Wallis, Mann-Whitney U, chi-square, Spearman correlation, multivariate logistic and linear regression.

Results: 15.2% of patients had NRS ≥ 3 (n=44). Median LOS in at-risk patients was significantly longer (12 vs 5 days; Kruskal-Wallis H=25.9, p<0.000001; Cohen's d=0.747). Gastro-Endoscopy patients showed higher risk than Gastro-Hepatology (22.8% vs 9.6%; χ²=8.57, p=0.003). Multivariate logistic regression identified two independent predictors: endoscopy ward (OR=2.37 [1.18–4.77]; p=0.016) and haemoglobin (OR=0.80 per g/dL [0.68–0.94]; p=0.007). NRS-2002/GLIM concordance showed high specificity (94.1%) but moderate sensitivity (45.6%).

Conclusion: NRS-2002 identifies a clinically relevant proportion of at-risk patients in Gastroenterology, with a strong impact on LOS. Haemoglobin emerges as an easily accessible biomarker and independent predictor of nutritional risk. Screening quality remains a critical issue. NRS-GLIM integration is feasible but requires systematic data collection.

Disclosure of Interest: None declared