These tools will no longer be maintained as of December 31, 2024. Archived website can be found here. PubMed4Hh GitHub repository can be found here. Contact NLM Customer Service if you have questions.


Pubmed for Handhelds

PUBMED FOR HANDHELDS

Search MEDLINE/PubMed

  • Title: GLIM criteria has fair sensitivity and specificity for diagnosing malnutrition when using SGA as comparator.
    Author: Allard JP, Keller H, Gramlich L, Jeejeebhoy KN, Laporte M, Duerksen DR.
    Journal: Clin Nutr; 2020 Sep; 39(9):2771-2777. PubMed ID: 31918864.
    Abstract:
    BACKGROUND & AIMS: The Global Leadership Initiative on Malnutrition (GLIM) proposed a new framework for diagnosing malnutrition based on combinations of phenotypic and etiologic criteria. The aim of this study was to compare GLIM criteria to Subjective Global Assessment (SGA) judged to be the most validated standardized assessment of malnutrition. METHODS: This is a retrospective analysis of variables extracted from a prospective cohort study assessing malnutrition at admission, in 18 Canadian hospitals. Based on the available parameters, GLIM was compared to SGA using the following combinations of one phenotypic and one etiologic criteria: A. weight loss and low intake; B. weight loss and high C-reactive protein (CRP); C. low body mass index (BMI) and low intake; D. low BMI, high CRP. Data were not available for fat-free mass. Since all patients had acute or chronic active disease as per GLIM etiologic criterion, CRP was used as a more specific measure to define inflammation. Sensitivity, specificity, positive (PPV) and negative (NPV) predictive values were calculated. Data are expressed as mean and Clopper-Pearson exact 95% confidence interval (CI). RESULTS: From 1022 patients in the original dataset, 784 had all considered parameters with a prevalence of malnutrition (SGA B or C) of 45.15% (CI 41.60, 48.70), where severe malnutrition (SGA C) was 11.73% (CI 9.57, 14.20). Using the available GLIM parameters with the above combinations of two-criteria, the prevalence of malnutrition was 33.29% (CI 30.00, 36.71) and severe malnutrition was 19.77% (CI 17.00, 22.70). For all criteria combinations of GLIM together versus SGA, sensitivity was 61.30% (CI 56.0, 66.4), specificity was 89.77% (CI 86.5, 92.5) and PPV was 83.14% (CI 78.0, 87.5) while NPV was 73.80 (CI 69.8, 77.5). Sensitivity was improved when only SGA C for severe malnutrition was used as the criterion (82.61%; CI 73.3, 89.7) but PPV was greatly reduced (29.12%; CI 23.7, 35.0). Similarly, when using GLIM criteria for severe malnutrition only, sensitivity improved (76.09%; CI 66.1, 84.4). Any two criteria combinations of GLIM had much poorer sensitivity with the highest being weight loss + high CRP (46.33%) with a specificity of 93.02% (PPV: 84.54%; NPV: 67.80%), while the combination of low BMI + low intake had the highest specificity (98.84%) but with a sensitivity of 15.54% (PPV 91.67%; NPV: 58.70%). CONCLUSIONS: Based on the CMTF dataset and using SGA as the most validated tool for diagnosing malnutrition, the two criteria combinations used for GLIM in the present study had fair criterion validity for the diagnosis of malnutrition, regardless of severity status. The best combinations were weight loss and high CRP or weight loss and low intake, both having high specificity at diagnosing malnutrition but unacceptably low sensitivity, and thus were considered poor. There may be potential for the full framework to be used to diagnose malnutrition, but individual combinations of two criteria when used exclusively will miss malnourished patients, as defined by SGA.
    [Abstract] [Full Text] [Related] [New Search]