Prevalence of disease related malnutrition assessed by NRS2002, STAMP, MNA-SF, one-step and two-step GLIM at admission and discharge: A systematic review and meta-analysis.

Abstract

Background and objectives: Disease-related malnutrition (DRM) mandates routine screening in many nations to ensure quality care as nutritional status often deteriorates during hospitalization. This study aimed to establish tool-specific prevalence benchmarks at admission and discharge, investigate sources of heterogeneity and evaluate diagnostic yields of one-step versus two-step GLIM approaches.

Methods and study design: A systematic search of Embase, PubMed, Web of Science, Cochrane Library, and CINAHL was conducted for cross-sectional studies on nutritional status published between 1 January 2018 and 30 Nov 2025 (PROS-PERO ID: CRD42023480467). Eligible studies were pooled in a meta-analysis of proportions using a random-effects model.

Results: Of 9,693 records retrieved, 179 studies involving 755,092 patients from 30 countries were included. At admission, the pooled prevalence was 0.41 (95% CI 0.35-0.49, I2 = 99.9%) by NRS2002 (≥3, adults), 0.62 (95% CI 0.57-0.67, I2 = 99.3%) by MNA-SF (≤11, older adults), 0.55 (95% CI 0.44-0.68, I2 = 98.5%) by STAMP (≥2, pediatric patients), and 0.39 (95% CI 0.35-0.44, I2 = 99.6%) by GLIM (adults). At discharge, three studies reported a pooled prevalence of 0.19 (95% CI 0.09-0.40, I2 = 94.6%) by NRS2002 for adults, significantly lower than admission (p = 0.044). Comparing GLIM strategies, no statistically significant difference was observed between one-step and two-step GLIM approaches (p = 0.062).

Conclusions: DRM prevalence remains high at both admission and discharge, varying widely across countries and by tools. Discharge data are critically limited, reflecting gaps in routine screening practices. Regarding diagnosis, findings support a risk-adapted strategy. Pooled estimates may overestimate the true prevalence and should be interpreted with caution.