Background and Objectives: This study validated the Malnutrition Universal Screening Tool (MUST) for nutritional screening in acutely unwell elderly patients against a reference assessment tool - Patient-Generated Subjective Global Assessment (PG-SGA). Methods and Study Design: One hundred and thirty two acutely admitted general medical patients contributed data for this study. In addition to performance of MUST and PG-SGA the following nutritional parameters were measured: weight loss > 5% in previous 3-6 months, handgrip strength, triceps skinfold thickness, Mid-arm circumference, Mid-arm muscle circumference (MAMC). Quality of life (QoL) was determined using the EuroQoL Questionnaire (EQ-5D 5 level). Sensitivity, specificity, predictive values and concordance were calculated to validate MUST against PG-SGA. Results: MUST when compared to PG-SGA gave a sensitivity of 69.7%, specificity of 75.8%, positive predictive value of 75.4%, negative predictive value of 70.1% and kappa statistics showed 72.7% agreement (k=0.49) for detecting malnutrition. The MUST score had significant inverse correlation with body mass index, Triceps skinfold thickness and Mid-arm muscle circumference but not with Handgrip strength. Malnourished patients (PG-SGA class B/C) were found to have a significantly worse QoL. Conclusions: This study demonstrates that MUST can be confidently administered with respect to validity in acutely unwell general medical elderly patients to detect malnutrition. In this study, significant recent weight loss also seems to have validity, almost comparable to MUST, for predicting the risk of malnutrition. Further research is needed to verify this finding, as a single item may be more feasible to complete than an instrument consisting of two or more items.