Valor prognóstico do GLIM para desfechos clínicos em pacientes com câncer colorretal e de pulmão

Abstract

Introduction and Objectives: The Global Leadership Initiative on Malnutrition (GLIM) was developed to standardize the diagnosis of malnutrition. While evidence indicates consensus in diagnosing malnutrition among oncology patients and its association with adverse clinical outcomes, methodological limitations underscore the need for further research, particularly in outpatient settings, which have been less explored. This study evaluated the prognostic value of various GLIM combinations, investigating their diagnostic accuracy and predictive value in outpatients with colorectal and lung cancer. Methods: 92 adult outpatients diagnosed with colorectal or lung cancer at various stages of treatment (pre-treatment or during pharmacological treatment) were assessed to determine the prognostic value of the GLIM tool. Data collection occurred between December 2022 and July 2024. The study was approved by the ethics committee of the Federal University of Minas Gerais (registration 91570318.5.0000.5149), and all participants provided informed consent. Patients were assessed prior to undergoing PET-CT at the Hermes Pardini Institute using a standardized questionnaire that gathered identification and clinical history, sociodemographic data, dietary intake, anthropometric and body composition measurements, muscle strength and functionality assessments, and laboratory markers. The Subjective Global Assessment (SGA) served as the reference tool for nutritional diagnosis. It classified patients as well-nourished, moderately malnourished, or severely malnourished. The GLIM diagnosis of malnutrition was based on the presence of at least one phenotypic criterion combined with one etiologic criterion. Eight distinct GLIM models were evaluated, with phenotypic variables kept constant. Etiologic variables included the presence of inflammation [identified by cancer and/or the modified Glasgow Prognostic Score (mGPS) and/or reduced food intake (assessed by a 24-hour recall and plate model)]. The different GLIM combinations were compared to the SGA for concurrent analysis. Primary outcomes for predictive analysis were obtained through telephone contact 12 months after the initial assessment, evaluating: (a) chemotherapy discontinuation (for any reason); (b) hospitalization need (for non-surgical causes); (c) mortality occurrence (all causes). Data were analyzed using SPSS v. 28.0, with a significance level of 5%. Sample size calculation for concurrent analysis was based on the presence of malnutrition, with a minimum of 66 patients required to achieve 80% statistical power. In the concurrent analysis, agreement between GLIM variables and SGA was assessed using the kappa test, interpreted as follows: ≤0.20, poor; 0.21–0.40, fair; 0.41–0.60, moderate; 0.61–0.80, good; and 0.81–1.00, very good. Sensitivity, specificity, positive and negative predictive values were assessed via cross-tabulation, with values above 80% considered satisfactory. The predictive capacity of GLIM for clinical outcomes was assessed using univariate logistic regressions adjusted for sex, age, tumor location, stage, and performance status. To compare malnutrition severity, the Kaplan-Meier survival method and log-rank test were employed. Results: The mean age of patients was 64.7 ± 1.3 years, with malnutrition prevalence ranging from 30.4% to 89.1% across different GLIM combinations. The GLIM combination that included the presence of cancer as an etiologic criterion demonstrated high sensitivity (100%), but low specificity (20%) and poor agreement (kappa 0.193), and was not independently associated with clinical outcomes. The GLIM combination that considered reduced food intake, identified by the adapted nDay Survey, showed moderate agreement with the Subjective Global Assessment (SGA) (kappa 0.446, accuracy 73%, specificity 86%) and predicted the risk of mortality and hospitalization within twelve months. Additionally, the combination utilizing the modified Glasgow Prognostic Score (mGPS) inflammatory marker exhibited high specificity (82%), highlighting its utility in confirming malnutrition, especially when associated with at least one phenotypic alteration in patients with colorectal and lung cancer. GLIM combinations 3–7 predicted 12-month mortality, whereas GLIM combinations 3–4 and GLIM 7 were independently associated with the risk of hospitalization. However, GLIM models were limited in classifying malnutrition severity of malnutrition, as survival rates were similar between moderately and severely malnourished patients. Conclusion: The malnutrition diagnosis based on the GLIM criteria demonstrated variable agreement with the Subjective Global Assessment (SGA). These findings suggest that, although objective methods for assessing dietary intake and inflammatory markers may not always be prioritized in nutritional status evaluations, recognizing these alterations in oncology patients improves the malnutrition diagnosis and highlights the need for early therapeutic interventions to achieve better clinical outcomes in this population. Further research is needed to incorporate methods for identifying reduced food intake and laboratory markers of inflammation, particularly those associated with oncologic disease, as variables in the GLIM criteria, thus enhancing the validation of this diagnostic tool.