Validity of bioelectrical impedance analysis in estimation of fat-free mass in colorectal cancer patients.

ABSTRACT

BACKGROUND & AIMS: Bioelectrical impedance analysis (BIA) is an accessible and cheap method to measure fat-free mass (FFM). However, BIA estimates are subject to uncertainty in patient populations with altered body composition and hydration. The aim of the current study was to validate a whole-body and a segmental BIA device against dual-energy X-ray absorptiometry (DXA) in colorectal cancer (CRC) patients, and to investigate the ability of different empiric equations for BIA to predict DXA FFM (FFMDXA). METHODS: Forty-three non-metastatic CRC patients (aged 50-80 years) were enrolled in this study. Whole-body and segmental BIA FFM estimates (FFMwhole-bodyBIA, FFMsegmentalBIA) were calculated using 14 empiric equations, including the equations from the manufacturers, before comparison to FFMDXA estimates. RESULTS: Strong linear relationships were observed between FFMBIA and FFMDXA estimates for all equations (R2 = 0.94-0.98 for both devices). However, there were large discrepancies in FFM estimates depending on the equations used with mean differences in the ranges -6.5-6.8 kg and -11.0-3.4 kg for whole-body and segmental BIA, respectively. For whole-body BIA, 77% of BIA derived FFM estimates were significantly different from FFMDXA, whereas for segmental BIA, 85% were significantly different. For whole-body BIA, the Schols* equation gave the highest agreement with FFMDXA with mean difference ±SD of -0.16 ± 1.94 kg (p = 0.582). The manufacturer's equation gave a small overestimation of FFM with 1.46 ± 2.16 kg (p < 0.001) with a tendency towards proportional bias (r = 0.28, p = 0.066). For segmental BIA, the Heitmann* equation gave the highest agreement with FFMDXA (0.17 ± 1.83 kg (p = 0.546)). Using the manufacturer's equation, no difference in FFM estimates was observed (-0.34 ± 2.06 kg (p = 0.292)), however, a clear proportional bias was detected (r = 0.69, p < 0.001). Both devices demonstrated acceptable ability to detect low FFM compared to DXA using the optimal equation. CONCLUSION: In a population of non-metastatic CRC patients, mostly consisting of Caucasian adults and with a wide range of body composition measures, both the whole-body BIA and segmental BIA device provide FFM estimates that are comparable to FFMDXA on a group level when the appropriate equations are applied. At the individual level (i.e. in clinical practice) BIA may be a valuable tool to identify patients with low FFM as part of a malnutrition diagnosis.