Predictive validity of GLIM malnutrition diagnosis in patients with colorectal cancer.

Background: Considering that the validation of the Global Leadership Initiative on Malnutrition (GLIM) remains unclear in patients with colorectal cancer, the present study aimed to assess the agreement, accuracy, sensitivity, specificity, and prognostic effect of the GLIM on survival when compared with the Patient-Generated Subjective Global Assessment (PG-SGA).

Methods: Patients with colorectal cancer who were scheduled to undergo a routine abdominal computed tomography (CT) scan were recruited. Using the GLIM two-step approach, the patients were first screened for malnutrition by using the PG-SGA Short Form (score ≥3). The malnutrition diagnosis was based on the etiologic (disease burden [cancer] or reduced food intake) and phenotypic GLIM criteria, including weight loss, body mass index, and skeletal muscle index at the third lumbar vertebra when using the CT scans. The food intake was assessed by the PG-SGA.

Results: This study included 191 patients (age, 60.5 ± 11.3 years; 57% men), and 23% and 32% were malnourished according to the GLIM and the PG-SGA, respectively. The GLIM revealed fair sensitivity (64%), good agreement (kappa = 0.65), specificity (96%), and diagnostic accuracy for detecting malnutrition (area under the receiver operating characteristic curve = 0.80; 95% CI, 0.72-0.88) when compared with the PG-SGA. The malnutrition value according to the GLIM and the PG-SGA was associated with short-term survival. However, only the PG-SGA was associated with long-term survival.

Conclusions: Although showing fair sensitivity, the GLIM had good agreement, specificity, and diagnostic accuracy for malnutrition detection and was an independent predictor of short-term survival in patients with colorectal cancer.