RESUMO
INTRODUCTION: Nutriscore is a malnutrition screening tool designed specifically for cancer patients. Our objective was to assess its performance in hospitalized cancer patients. PATIENTS AND METHODS: Adult patients diagnosed with any solid neoplasm hospitalized in Medical Oncology were included. In the first 24-48 h, of admission they were screened with Nutriscore and Malnutrition universal screening tool (MUST). Both tests were compared using chi-square, kappa index and ROC curve. Nutriscore sensitivity (S), specificity (Sp) and predictive values (PV) were calculated using MUST as a reference. RESULTS: A total of 93 patients were included. The most frequent tumors were lung (36.6%), colorectal (24.8%) and breast (8.6%). MUST identified 69.9% of the patients at nutritional risk, and Nutriscore 44.1% (p < 0.001), with a low kappa index [k = 0.38 (95% CI 0.23 to 0.54)]. The AUC of the ROC curve for Nutriscore with respect to the MUST was 0.739. Nutriscore showed S = 58.6 (95% CI 45.7 to 71.2), Sp = 89.3% (95% CI 76.0 to 100.0%), VP + = 92.7% (95% CI 83.5 at 100.0%) and VP- = 48.1% (95% CI 33.5 to 62.6). CONCLUSIONS: Nutriscore did not provided better screening results in hospitalized cancer patients than a validated tool such as MUST.