• Paula Von Geldern, Claudio Salas, Pablo Alvayay, Claudio Silva, and Maria Pía de la Maza.
• School of Nutrition and Dietetics, Mayor University, Santiago, Chile; Institute of Nutrition & Food Technology Dr. Fernando Monckeberg Barros (INTA), University of Chile, Santiago, Chile.
• Nutrition. 2021 Apr 1; 84: 111006.

ObjectivesThe aim of this study was to analyze the association between survival and two validated methods of nutritional assessment: body composition through computed tomography (CT) scans and Patient-Generated Subjective Global Assessment (PG-SGA).MethodsCancer-bearing patients (n = 103) hospitalized in the Oncology Ward of Clínica Alemana in Santiago, Chile, for palliative or curative treatment were assessed by both methods. Images from abdominal CT scans at the L3 level were analyzed by SliceOmatic (version 5.0), to measure muscle and fat areas and densities. Skeletal muscle mass index (MMI) was calculated using total abdominal mass area (psoas + rest of muscles)/ height2. These were compared with those obtained for assessment of trauma of 130 healthy young adults (18-40 y of age), as reference control values. Sarcopenia was established as MMI<1 SD compared with control participants.ResultsPatients with cancer had less muscle and higher abdominal fat areas compared with controls (P < 0.05). According to the PG-SGA, ~50% were classified as malnourished. Patients were followed for 38 mo, when 53% had died. Survival time was significantly and negatively correlated with PG-SGA score, cancer stage, and sarcopenia, independent of age and sex. Multivariate analysis included both cancer stage and nutritional assessment variables.ConclusionsTogether with cancer stage, both CT measurements and subjective assessment of nutritional status through PG-SGA can adequately identify cancer patients with a higher mortality risk, independent of age and sex. However, the latter is less costly and simple to use; it should be included as a valuable tool during management of patients with cancer.Copyright © 2020 Elsevier Inc. All rights reserved.

5 keywords...

hide…