Prevalence of malnutrition comparing NRS2002, MUST, and PG-SGA with the GLIM criteria in adults with cancer: A multi-center study

Objectives: This study aimed to evaluate the diagnostic capacity of the Nutritional Risk Screening 2002 (NRS2002), Malnutrition Universal Screening Tool (MUST), and Patient-generated Subjective Global Assessment (PG-SGA) in light of the Global Leader Initiative onMalnutrition (GLIM) criteria in adult patients with cancer. Methods: A multicenter observational study was conducted. Nutritional screening and assessment were performed at the time of admission to hospitals with the NRS2002, MUST, PG-SGA, and GLIM criteria. Sensitivity, specificity, positive and negative predictive values, positive and negative likelihood ratio, and Kappa (K) values were used to evaluate the performance of these tools. Results: Of the 637 included patients, 24.8% and 15.4% of patients were at moderate and high risk of malnutrition, respectively, using the NRS2002 and MUST. The NRS2002 was better correlated to the GLIM criteria with a higher value of Kappa (K = 0.823 vs. 0.596) and area under the receiver operating characteristic curve (K = 0.896 vs. 0.757) than the MUST. Meanwhile, 28.3% of patients were diagnosed as malnourished at the time of admission per the GLIM criteria, and 43.3% were malnourished per the PG-SGA. The PG-SGA had a fair agreement with the GLIM criteria (K = 0.453), revealing a positive predictive value of 52.9% and negative predictive value of 90.6%. Conclusions: The NRS2002 was better correlated with the GLIM diagnostic criteria of malnutrition than the MUST. The PG-SGA was too sensitive to detect nutrition-related deteriorations, leading to a low positive predictive value in the malnutrition diagnosis. Thus, the GLIM criteria could be used to confirm the presence of malnutrition identified by the PG-SGA in adults with cancer. (C) 2020 Elsevier Inc. All rights reserved.

Automated summary

The study found that NRS2002 was better correlated with the GLIM diagnostic criteria of malnutrition than MUST, while PG-SGA was too sensitive in detecting nutrition-related deteriorations, resulting in a lower positive predictive value in malnutrition diagnosis.