A comparative study of the Global Leadership Initiative on Malnutrition criteria vs the Patient-Generated Subjective Global Assessment in predicting liver cirrhosis survival

Abstract

Objective — to study the impact of malnutrition, determined by the Global Leadership Initiative on Malnutrition (GLIM) criteria and the Patient-Generated Subjective Global Assessment (PG-SGA), on the long-term survival of liver cirrhosis (LC) patients and to compare the prognostic value of these tools. Materials and methods. A cohort of 170 patients with LC of viral, alcoholic, or mixed aetiology (58 women and 112 men, mean age 55.9 ± 11.2 years) was enrolled in the prospective study. The median follow-up period was 490 (IQR 293 — 642) days. During this time, 53 patients died due to liver-related complications. Nutritional state was assessed using the GLIM criteria and PG-SGA. The skeletal muscle index was utilised to assess muscle mass reduction, as a phenotypic GLIM criterion. The presence of inflammation, as an aetiologic criterion, was defined as a C-reactive protein level above 5 mg/L. Results. Despite the similar prevalence of malnutrition detected using the GLIM criteria and PG-SGA (70.6 % and 61.8 % of patients, respectively), the diagnostic concordance between instruments was low (k Cohen’s = 0.518). The severity of liver cirrhosis is directly correlated with the prevalence of malnutrition. Over 85 % of patients with Child–Turcotte–Pugh Class C cirrhosis were diagnosed with impaired nutritional state by either the GLIM criteria or PG-SGA. The GLIM criteria demonstrated poor predictive value for mortality (AUC 0.625, p = 0.006), whereas the PG-SGA showed acceptable predictive value (AUC 0.703, p = 0.000). Kaplan–Meier survival analysis revealed that the mortality rate among patients with impaired nutritional state, as assessed by both GLIM criteria and PG-SGA, was significantly higher than in those with normal nutritional state. The greatest negative impact of malnutrition on survival is concentrated within the first 6 months of follow-up. In the multivariate analysis only hypoalbuminemia, hepatic encephalopathy, and malnutrition determined by the PG-SGA had an independent impact on the time to death. The hazard ratio (HR) for mortality associated with malnutrition according to PG-SGA was 2.665 (p = 0.027). Malnutrition diagnosed by GLIM criteria was not an independent predictor of mortality (HR 1.617, p = 0.304).