Background & aims: Malnutrition is prevalent among hospitalized patients, but there is no universally accepted consensus regarding its diagnosis. Recently, the Global Leadership Initiative on Malnutrition (GLIM) proposed a new framework for the malnutrition diagnosis and until this moment there is scarce evidence regarding its validity. This study aimed to evaluate the concurrent and predictive validity of GLIM criteria for malnutrition diagnosis in hospitalized patients.
Methods: Prospective cohort study involving adult/elderly hospitalized patients. The malnutrition diagnoses according to Subjective Global Assessment (SGA) and GLIM criteria were performed within 48 h of admission. Patients were followed up until hospital discharge to assess the length of hospital stay (LOS) and in-hospital mortality. Six months post discharge; the patients were contacted to collect the outcomes readmission and death. Agreement and accuracy tests, Cox and Logistic regression analysis were performed for testing criterion validity.
Results: 601 patients (55.7 ± 14.8 years, 51.3% men) were evaluated. Malnutrition was diagnosed in 33.9% and 41.6% of patients, by SGA and GLIM criteria, respectively. GLIM criteria presented a satisfactory accuracy, (AUC = 0.842; CI95% 0.807-0.877) with a sensitivity of 86.6%, and a specificity of 81.6%. The presence of malnutrition by GLIM criteria increased the chance of prolonged hospitalization by 1.76 (CI95% 1.23-2.52) times, and the risk of in-hospital deaths by 5.1 (CI95% 1.14-23.14) times. It was also associated with death within six months (RR = 3.96, CI95% 1.49-10.53).
Conclusion: GLIM criteria for malnutrition diagnosis presented satisfactory criterion validity and should be applied during clinical practice.
Keywords: Hospital; Length of stay; Malnutrition; Mortality; Nutrition assessment.
Copyright © 2021. Published by Elsevier Ltd.