MALNUTRITION IN PATIENTS WITH GASTROINTESTINAL CANCER: EFFECTIVENESS OF DIFFERENT DIAGNOSTIC METHODS

to estimate the effectiveness of methods for identifying the risk and/or presence of malnutrition in individuals with gastrointestinal neoplasia.

participated 143 patients with gastrointestinal cancer, cared for in the Oncology Clinic Infirmary of "Hospital Amaral Carvalho" (Jaú-SP). Excluded from the study were patients hospitalized in the intensive care unit, in a terminal state or those who had members amputated; those who had received blood transfusions during the last month; significant clinical bleeding; received endovenous albumin and uncontrolled infection. The nutritional status was classified according to the ratio between Real Weight and Habitual Weight, Body Mass Index, Nutritional Risk Index and Percent Adequacy. As the gold standard method the Scored Patient-Generated Subjective Global Assessment was used. The effectiveness of the methods for detecting the risk for or presence of malnutrition was evaluated. A ROC curve was constructed and its area (AUROC) was estimated. The areas were compared using z statistics. For each method the best cut-off point was established.

of the patients, 74.1% presented an advanced stage of the disease, and 83.2% were undergoing chemotherapy treatment. All the methods showed adequate discriminatory capacity for detecting the risk of malnutrition and presence of malnutrition. The BMI was significantly better for detecting malnutrition than for the risk of malnutrition. The RW/HW was significantly better for detecting the risk of malnutrition than the other methods. The cut-off points were slightly lower than those recommended for the normative population for the methods RW/HW, NRI and Score %. For the BMI the cut-off point was higher than the recommended for the normative population.

the methods for evaluating nutritional status showed adequate discriminatory capacity for the risk of malnutrition and presence of malnutrition in patients with gastrointestinal cancer.