RESUMO
BACKGROUNDS & AIMS: Malnutrition is prevalent among hospitalized patients in developed countries, contributing to negative health outcomes and increased healthcare costs. Timely identification and management of malnutrition are crucial. The lack of a universally accepted definition and standardized diagnostic criteria for malnutrition has led to the development of various screening tools, each with varying validity. This complicates early identification of malnutrition, hindering effective intervention strategies. This systematic review and meta-analysis aimed to identify the most valid and reliable nutritional screening tool for assessing the risk of malnutrition in hospitalized adults. METHODS: A systematic literature search was conducted to identify validation studies published from inception to November 2023, in the Pubmed/MEDLINE, Embase, and CINAHL databases. This systematic review was registered in INPLASY (INPLASY202090028). The risk of bias and quality of included studies were assessed using the Quality Assessment of Diagnostic Accuracy Studies version 2 (QUADAS-2). Meta-analyses were performed for screening tools accuracy using the symmetric hierarchical summary receiver operative characteristics models. RESULTS: Of the 1646 articles retrieved, 60 met the inclusion criteria and were included in the systematic review, and 21 were included in the meta-analysis. A total of 51 malnutrition risk screening tools and 9 reference standards were identified. The meta-analyses assessed four common malnutrition risk screening tools against two reference standards (Subjective Global Assessment [SGA] and European Society for Clinical Nutrition and Metabolism [ESPEN] criteria). The Malnutrition Universal Screening Tool (MUST) vs SGA had a sensitivity (95% Confidence Interval) of 0.84 (0.73-0.91), and specificity of 0.85 (0.75-0.91). The MUST vs ESPEN had a sensitivity of 0.97 (0.53-0.99) and specificity of 0.80 (0.50-0.94). The Malnutrition Screening Tool (MST) vs SGA had a sensitivity of 0.81 (0.67-0.90) and specificity of 0.79 (0.72-0.74). The Mini Nutritional Assessment-Short Form (MNA-SF) vs ESPEN had a sensitivity of 0.99 (0.41-0.99) and specificity of 0.60 (0.45-0.73). The Nutrition Universal Screening Tool-2002 (NRS-2002) vs SGA had a sensitivity of 0.76 (0.58-0.87) and specificity of 0.86 (0.76-0.93). CONCLUSIONS: The MUST demonstrated high accuracy in detecting malnutrition risk in hospitalized adults. However, the quality of the studies included varied greatly, possibly introducing bias in the results. Future research should compare tools within a specific patient population using a valid and universal gold standard to ensure improved patient care and outcomes.