• Micheline Tereza Pires Souza, Pierre Singer, Gislaine Aparecida Ozorio, Vitor Modesto Rosa, Maria Manuela Ferreira Alves, Rossana Verónica Mendoza López, and Dan L Waitzberg.
• Indirect Calorimetry and Body Composition Clinic of the Cancer Institute of the State of São Paulo, São Paulo, Brazil. Electronic address: Micheline.pires@gmail.com.
• Nutrition. 2018 Jul 1; 51-52: 60-65.

ObjectivesPatients with head and neck cancer have changes in body composition and resting energy expenditure (REE) related to significant inflammatory processes. We investigated REE and body composition in a population of patients with head and neck cancer, comparing the measured REE with predicted energy expenditure and deriving an equation of anthropometric values and body composition.MethodsThis retrospective, observational, descriptive study of a single center included patients with head and neck cancer. We evaluated nutritional status by body mass index (BMI) and Patient-Generated Subjective Global Assessment (PG-SGA), body composition by electric bioimpedance, and REE by indirect calorimetry (IC).ResultsWe included 140 patients, most of whom were men (80.7%), 60 y or older (58.6%), and had advanced disease (77.9%). Most were malnourished by BMI standards (77.9%) and severely malnourished according to the PG-SGA (49.3%), with a fat-free mass below the ideal values (82.9%) associated with sarcopenia (92.1%). Hypermetabolism was 57%. When comparing REE with the Harris-Benedict formula, we found the agreement limits from -546 613 to 240 708, the mean difference was -152 953 (95% confidence interval [CI], -185 844 to -120 062) and Pitman's variance test was r = -0.294 (P = 0.001). When we included the activity factor and the thermogenesis factor in REE and compared with Harris-Benedict, we found the agreement limits from -764.423 to 337.087, a mean difference of -213.668 (95% CI -259.684 to -167.652), and the Pitman's variance text at r = -0.292 (P = 0.001).ConclusionPredictive equations, generally recommended by guidelines, are imprecise when compared with IC measures. Therefore, we suggest a new predictive equation.Copyright © 2018 Elsevier Inc. All rights reserved.

5 keywords...

hide…