RESUMO
BACKGROUND: Cancer is a disease with high and increasing incidence rates in the world and its course tends to harm the body composition. Monitoring these body changes is very important. Therefore, it is essential to have reliable, accessible, and practical methods for evaluating body compartments. This study aims to evaluate the correlation and agreement of results for the bioelectrical impedance analysis (BIA) obtained from devices from different manufacturers. METHODS: This is a single-center cross-sectional study including hospitalized patients with cancer. Two devices from different brands used for obtaining the BIA were used; both with a tetrapolar model and a single frequency (50 kHz). The results were evaluated for resistance (R) and reactance (Xc) and used to calculate the phase angle (PhA) and fat-free mass (FFM) indicators. Pearson and Spearman correlation tests and Bland-Altman plots were performed, with results expressed as bias and limits of agreement at 95% confidence intervals (95%CI). RESULTS: We have included 116 patients, with a mean age of 60.8 ± 14.8, 51.7% were women. We have found very strong correlations between the measurements of R (rho = 0.971) and FFM (r = 0.979), and strong correlations for Xc (rho = 0.784) and PhA (rho = 0.768). However, the measurements did not agree between the methods. CONCLUSIONS: Commercial brands of devices used for the BIA influence the results generated, a factor that must be considered when choosing the most appropriate method for this analysis.