Comparative Clinical Efficacy of the Global Leadership Initiative on Malnutrition (GLIM) and the Patient-Generated Subjective Global Assessment (PG-SGA) in Gastric Cancer Patients

Objective: This study aims to compare the effectiveness of the Global Leadership Initiative on Malnutrition (GLIM) and the Patient-Generated Subjective Global Assessment (PG-SGA) in diagnosing malnutrition and predicting survival in gastric cancer patients. Additionally, it seeks to evaluate different muscle measurement indices used in GLIM criteria to determine their association with survival and refine the selection of sarcopenia assessment indicators. Methods: This multicenter, prospective cohort study involved 1,295 gastric cancer patients. We calculated the 5th and 15th percentiles for mid-arm circumference (MMC), calf circumference (CC), and handgrip strength index adjusted for weight (HGS/W).Various muscle measurement indices were employed to formulate the GLIM criteria and determine the optimal cut-off values for the diagnosis of sarcopenia, with an analysis focused on their capability to differentiate survival outcomes, ultimately assessing their prognostic evaluation potential. Both PG-SGA and GLIM were employed to diagnose and stage malnutrition, and their reliability and validity were compared. The correlation of malnutrition diagnoses from GLIM and PG-SGA with hospitalization costs and clinical blood test indicators was assessed to evaluate the clinical utility of these diagnostic methods. Results: Severe malnutrition diagnosed by GLIM, regardless of whether MMC, CC, or HGS/W was used as the positive criterion for sarcopenia, was associated with the shortest median overall survival (OS) and the highest hazard ratio (HR=1.563, 95% CI=1.314-1.860) compared to the normal group. GLIM demonstrated higher sensitivity but lower specificity compared to PG-SGA. Both methods indicated that greater malnutrition severity was linked to an increased risk of death. Cox regression analysis identified staging, KPS score, and PG-SGA-based malnutrition grading as independent survival risk factors. Malnutrition identified by both methods was significantly correlated with hospitalization costs, NRS2002 score, KPS score, hemoglobin, albumin, and creatinine (P<0.05), suggesting substantial clinical predictive value. Conclusion: While GLIM shows certain sensitivity in diagnosing malnutrition relative to PG-SGA, its specificity is relatively lower. Both diagnostic methods are valuable for predicting clinical outcomes, with PG-SGA demonstrating stronger predictive power for survival risk.