Scored-GLIM as an effective tool to assess nutrition status and predict survival in patients with cancer

• DOI: 10.1016/j.clnu.2021.01.033
• Publication Date: 2021-01-29T09:23:02Z
• Authors: Qi Zhang; Kang-ping Zhang; Xi Zhang; Meng Tang; Chun-hua Song; Ming-hua Cong; Zeng-qing Guo; Jia-shan Ding; Marco Braga; Tommy Cederholm; Hong-xia Xu; Wei Li; Rocco Barazzoni; Han-ping Shi
• Keywords: Male; Cancer patient; GLIM criteria; Malnutrition; Score; Survival;; Survival; Nutritional Status; Kaplan-Meier Estimate; Severity of Illness Index; 03 medical and health sciences; 0302 clinical medicine; Predictive Value of Tests; Reference Values; Neoplasms; Weight Loss; Humans; Cancer patient; GLIM criteria; Malnutrition; Score; Survival; Aged; Proportional Hazards Models; Retrospective Studies; 2. Zero hunger; GLIM criteria; Malnutrition; Score; Middle Aged; Prognosis; 3. Good health; Nomograms; Nutrition Assessment; ROC Curve; Cancer patient; Female

The Global Leadership Initiative on Malnutrition (GLIM) released new universal criteria for diagnosing and grading malnutrition, and calls for further investigations not only in different clinical setting but also in GLIM itself including reference value, combination and weight of different GLIM criteria. This study aimed to weigh the GLIM criteria and develop a scored-GLIM system, and then validate as well as evaluate its application in nutritional assessment and survival prediction for patients with cancer.A total of 3547 patients in the primary cohort and 415 patients in the validation cohort were included in the study. Patients' nutritional status were retrospectively assessed using the GLIM criteria. Kaplan-Meier survival curves and multivariate Cox regression analyses were performed to analyze the association between nutritional status and overall survival (OS). A nomogram was produced to quantify the GLIM criteria and develop the scored-GLIM system. C-index, receiver operating characteristic (ROC) curve and calibration curve analyses were performed to validate the predictive accuracy and discriminatory capacity of the scored-GLIM. Finally, a decision curve was applied to assess the clinical utility of the scored-GLIM system.In the primary cohort, 70.3% of patients were diagnosed as malnutrition. The malnutrition severity grading according to the GLIM criteria were associated with the prognosis of patients with cancer (HR 1.42, 1.23 to 1.65 for moderate malnutrition; HR 1.80,1.84 to 2.09 for severe malnutrition). The weight of each GLIM criteria was calculated, and unintentional weight loss was the most determining factor acting upon mortality (HR 1.82, 1.64 to 2.10 for stage II and HR 1.50, 1.31 to 1.73 for stage I). A nomogram was constructed by four factors of GLIM to weigh the GLIM criteria. The areas under the ROC curve were 65.3 (1-year ROC) and 65.5 (3-year ROC), and the C-index was 0.62, and the calibration curves fitted well. Decision curve analysis demonstrated the clinical usefulness of the scored-GLIM system.The accuracy and net clinical benefit of scored-GLIM system were similar to scored-PG-SGA but higher than GLIM both in nutrition assessment and in survival prediction for patients with cancer. These findings, along with its time-savings advantages over scored-PG-SGA, suggest scored-GLIM be a better nutritional assessment tool.