Analysis of ESPEN and GLIM algorithms reveals specific drivers for the diagnosis of malnutrition in patients with chronic gastrointestinal diseases

Karen Bannert; Lea Franziska Sautter; Mats Lukas Wiese; Fatuma Meyer; Luise Ehlers; Sophie Fromhold-Treu; Cathleen Karbe; Simone Gärtner; Markus M Lerch; Ali A Aghdassi; Robert Jaster; Luzia Valentini; Georg Lamprecht

doi:10.1016/j.nut.2022.111887

Analysis of ESPEN and GLIM algorithms reveals specific drivers for the diagnosis of malnutrition in patients with chronic gastrointestinal diseases

Nutrition. 2023 Feb:106:111887. doi: 10.1016/j.nut.2022.111887. Epub 2022 Nov 1.

Affiliations

¹ Department of Medicine II, Division of Gastroenterology and Endocrinology, Rostock University Medical Center, Rostock, Germany.
² Department of Medicine A, University Medicine Greifswald, Greifswald, Germany.
³ University of Applied Sciences Neubrandenburg, Institute for Evidence-Based Dietetics, Neubrandenburg, Germany.
⁴ Department of Medicine II, Division of Gastroenterology and Endocrinology, Rostock University Medical Center, Rostock, Germany. Electronic address: georg.lamprecht@med.uni-rostock.de.

PMID: 36473417
DOI: 10.1016/j.nut.2022.111887

Abstract

Objectives: Disease-related malnutrition (MN) is common in patients with liver cirrhosis (LC), short bowel syndrome (SBS), and chronic pancreatitis (CP). Different MN risk screening tools and diagnostic criteria of the European Society for Clinical Nutrition and Metabolism (ESPEN) and Global Leadership Initiative on Malnutrition (GLIM) algorithms were analyzed for their diagnostic accuracy and role as specific drivers to diagnose MN in patients with LC, SBS, and CP.

Methods: A total of 187 patients with LC, SBS, and CP, as well as control patients were prospectively recruited in a multicenter cross-sectional study. MN risk was screened using Nutritional Risk Screening 2002 (NRS-2002), the Malnutrition Universal Screening Tool (MUST), and the Royal Free Hospital Nutritional Prioritizing Tool (RFH-NPT), and diagnosed using the ESPEN, GLIM, and GLIM^CRP+ (GLIM incorporating C-reactive protein [CRP] >5 mg/L) algorithms. For each of the individual diagnostic criteria, relative frequency, sensitivity, specificity, as well as positive and negative predictive values were calculated.

Results: NRS-2002 was only sensitive in conjunction with ESPEN, while MUST was sensitive additionally with the GLIM algorithm. RFH-NPT worked the best for LC. GLIM and GLIM^CRP+ diagnosed MN more frequently than the ESPEN algorithm. Diagnostic criteria were detected at remarkably different relative frequencies starting with reduced food intake/malabsorption and chronic disease/inflammation, followed by weight loss, reduced fat-free mass index, low body mass index, and body mass index <18.5 kg/m². Relative frequencies differed between LC, SBS, and CP. Weight loss in LC and CP and reduced fat-free mass index and food intake in SBS had good diagnostic accuracy, suggesting that these criteria act as specific drivers for MN.

Conclusions: RFH-NPT and MUST performed better in conjunction with the GLIM algorithm than NRS-2002. MN was diagnosed more frequently by GLIM than the ESPEN algorithm in LC, SBS, and CP. Individual criteria acted as specific drivers in MN in chronic gastrointestinal diseases.

Keywords: Chronic pancreatitis; ESPEN; GLIM; Liver cirrhosis; Malnutrition; Short bowel syndrome.

Publication types

Multicenter Study
Research Support, Non-U.S. Gov't

MeSH terms

Algorithms
Chronic Disease
Cross-Sectional Studies
Gastrointestinal Diseases* / complications
Gastrointestinal Diseases* / diagnosis
Humans
Liver Cirrhosis
Malnutrition* / diagnosis
Malnutrition* / etiology
Nutrition Assessment
Nutritional Status
Weight Loss