Comparison of body composition techniques in Portuguese patients with gastrointestinal cancer

Abstract

Introduction: Anthropometric measures and bioelectrical impedance have been used in the past for body composition assessment of cancer patients. However, more specific and precise methodologies, such as CT imaging, are now available. So the purpose of this study was to compare different body composition techniques, using CT scan as golden standard.

Materials and methods: Cross-sectional study with 100 patients with gastrointestinal cancer (GI) enrolled consecutively at diagnosis. Clinical data including disease site and stage were collected. Performance status was assessed with the Eastern Cooperative Oncology Group Performance Status scale. Nutritional assessment was conducted with anthropometric measures, namely weight, height, triceps skinfold, arm and waist circumference were recorded. Also, fat free mass, body fat, total body water, phase angle, as well as raw data as resistance and reactance were obtained with bioelectrical impedance analysis (BIA) (Bodystat 1500 MDD). Body composition was also assessed with Computed Tomography (CT), images were acquired at the 3rd lumbar vertebrae-L3, muscle and fat tissue cross sectional areas (cm²) and mean muscle attenuation (Hounsfield Units, HU) were recorded. Skeletal Muscular Index (SMI-cm²/height²) was calculated and gender and Body Mass Index (BMI) specific cut-offs for SMI were used to define sarcopenia [1]. This study was approved by the ethics committee of Hospital Beatriz Ângelo.

Results: Sarcopenia prevalence was 32% (n = 32). Sarcopenic patients were older (p = 0.003) and had a worse performance status (p = 0.06). Sarcopenic patients had a higher mean weight loss (p = 0.05), and a lower mean BMI (p = 0.03) although 56% of them were overweight/obese. CT defined sarcopenia could not be detected by many commonly used anthropometric measures, since no differences between sarcopenic and non-sarcopenic patients were found for waist circumference, triceps skin fold, arm muscle and fat area. When sarcopenic and non-sarcopenic patients were compared according to body composition data from BIA, the only significant differences refer to higher mean body water (p = 0.04) and lower phase angle (p = 0.04) in sarcopenic patients. Considering CT imaging as the gold standard, total fat area correlated with BMI (r = 0.862, p < 0.001) and waist circumference (r = 0.852, p < 0.001); subcutaneous fat area with FM from BIA (r = 0.810, p < 0.001) and visceral fat area with waist circumference (r = 0.782, p < 0.001). Skeletal muscle area from CT imaging was correlated with fat free mass from BIA (r = 0.745, p < 0.001). However, in absolute values no agreement was found between estimated whole body skeletal muscle using CT and BIA. Either weak or non-significant correlations were mostly observed between muscle radiation attenuation and anthropometry/BIA.

Discussion and conclusions: Although correlations between CT scan and anthropometry/BIA were found, body composition assessment with anthropometry and BIA may be insufficient to identify sarcopenia and fat infiltration of skeletal muscle tissue. Therefore CT scan analysis should be routinely used in the nutritional assessment of GI cancer patients.