Background: Early assessment of malnutrition in cancer patients is essential. This study analyzed the accuracy of the Global Leadership Initiative on Malnutrition (GLIM) and the Subjective Global Assessment (SGA), in diagnosing malnutrition, considering the Patient Generated-SGA (PG-SGA) as a reference, and the impact of malnutrition on hospital days.
Methods: We conducted a prospective cohort study in 183 patients with gastrointestinal, head and neck, and lung cancer. Malnutrition was assessed within 48 h, of hospital admission according to the SGA, PG-SGA, and GLIM. Accuracy tests and regression analysis were performed to assess the criterion validity of the GLIM and SGA for diagnosing malnutrition.
Results: Malnutrition was diagnosed in 57.3% (SGA), 86.3% (PG-SGA), and 74.9% (GLIM) of the inpatients. The median of hospitalization was 6 (3-11) days, with 47% hospitalized > 6 day. The SGA presented the best accuracy (AUC = 0.832) than the GLIM (AUC = 0.632) compared to PG-SGA. Patients diagnosed with malnutrition by SGA, GLIM, and PG-SGA remained hospitalized for 2.13, 3.19, and 4.56 day more than well-nourished patients, respectively.
Conclusion: Compared to PG-SGA, the SGA presents good accuracy and adequate specificity (>80%). Malnutrition evaluated by SGA, PG-SGA, and GLIM was associated with more days of hospitalization.