Abstract
OBJECTIVES: Global Leadership Initiative on Malnutrition (GLIM) criteria have gradually accounted for the mainstay evaluating nutritional status. We sought to establish GLIM-dictated nomograms with other prognostic factors influencing long-term mortality and externally validate their predictive performance in decompensated cirrhosis. METHODS: The derivation cohort comprised 301 patients presenting with cirrhosis-associated acute insults, while the validation cohort encompassed 101 subjects from another tertiary hospital. Two nomograms were constructed to predict the 1-year all-cause mortality by integrating the GLIM criteria. The study population was stratified into low-, moderate- and high-risk mortality groups according to aforesaid proposed models. RESULTS: Adjusting Child-Turcotte-Pugh classification (Nomo#1) or Model for End-stage Liver Disease-Sodium score (Nomo#2) separately, the GLIM criteria were independently associated with 1-year mortality in the multivariate Cox regression analysis (Nomo#1 hazard ratio (HR) = 3.139, p < 0.001; Nomo#2 HR = 3.456, p < 0.001). The C-index and time AUC for Nomo#1 and Nomo#2 performed significantly better than those of the GLIM criteria or conventional scoring systems alone. The survival rate of the low-risk group was significantly higher than those of the moderate- or high-risk groups (Nomo#1: 95% vs 65.8% vs 33.3%, p < 0.001; Nomo#2: 94.3% vs 64.5% vs 25%, p < 0.001). Furthermore, our proposed models exhibited moderate prediction accuracy and may identify malnourished patients with poor survival conditions in the external validation cohort. CONCLUSION: GLIM criteria-defined malnutrition negatively impacted long-term mortality in the context of decompensated cirrhosis. Our established nomograms may predict survival status with sufficient discriminatory ability, alongside good consistency and clinical benefits, supporting their effectiveness in daily practice.