The association between GLIM criteria–defined malnutrition and 2‐year unplanned hospital admission in outpatients with unintentional weight loss: A retrospective cohort study

Abstract Background This study aimed to assess malnutrition using the Global Leadership Initiative on Malnutrition (GLIM) criteria and Subjective Global Assessment (SGA) at baseline and determine the GLIM criteria that best predicted unplanned hospitalization in outpatients with unintentional weight loss (UWL). Methods We performed a retrospective cohort study of 257 adult outpatients with UWL. The GLIM criteria and SGA agreement were reported using the Cohen kappa coefficient. Kaplan‐Meier survival curves and adjusted Cox regression analyses were used for survival data. Logistic regression was used for the other correlation analysis. Results This study collected data from 257 patients for 2 years. Based on the GLIM criteria and SGA, malnutrition prevalence was 79.0% and 72.0%, respectively ( κ = 0.728, P < 0.001). Using the SGA as a standard, GLIM had a sensitivity of 97.8%, a specificity of 69.4%, a positive predictive value of 89.2%, and a negative predictive value of 92.6%. Malnutrition was associated with higher rates of unplanned hospital admission independent of other prognostic factors (GLIM: hazard ratio [HR]=2.85, 95% CI=1.22–6.68; SGA: HR=2.07, 95% CI=1.13–3.79). Of the five GLIM criteria–related diagnostic combinations, disease burden or inflammation was the most important to predict unplanned hospital admission in multivariable analysis (HR=3.27, 95% CI=2.03–5.28). Conclusion There was good agreement between the GLIM criteria and the SGA. GLIM‐defined malnutrition, as well as all five GLIM criteria–related diagnosis combinations, had the potential to predict unplanned hospital admissions in outpatients with UWL within 2 years.