Comparison of DEXA and Bioimpedance for Body Composition Measurements in Nondialysis Patients With CKD

Objectives: The aims of this study are (1) to compare dual-energy X-ray absorptiometry (DEXA) and bioimpedance for body composition measurements in nondialysis patients with chronic kidney disease, and (2) to investigate factors associated with any measurement differences. Design and Methods: This is a substudy using some baseline data from a randomized controlled clinical trial. One hundred twenty patients (aged 65 +/- 14 years) with a measured glomerular filtration rate 8 to 55 mL/min/1.73m(2), not on renal replacement therapy, irrespective of age and number of comorbidities, were included from 2011 to 2016. For DEXA measurements, Lunar Prodigy or Lunar iDXA were used. For bioimpedance measurements, body composition monitor (BCM) was used. Glomerular filtration rate was measured with iohexol clearance. Data were analyzed using R software. Bland-Altman analysis was performed to compare the 2 measurements. The measurement difference was DEXA minus BCM. Multiple linear regression analysis was performed to analyze relationships between variables. Results: The estimation of fat-free mass was higher using BCM than DEXA, with a mean difference of -2.8 kg and limits of agreement (mean +/- 2 SD) ranging from -12 kg to 6.5 kg. The estimation of fat mass was lower using BCM than DEXA, with a mean difference of 3.1 kg and limits of agreement (mean +/- 2 SD) ranging from -6.8 kg to 13 kg. The measurement differences were significantly related to lean tissue index, fat tissue index, extracellular water, intracellular water, extracellular water/intracellular water, total body water, and overhydration. Conclusion: Our study showed a limited agreement between DEXA and bioimpedance, indicating that these 2 measurements are not interchangeable in nondialysis patients with chronic kidney disease. Lean tissue index, fat tissue index, and body water might contribute to the measurement differences, while measured glomerular filtration rate is not a factor associated with the measurement differences for body composition. Thus, we suggest that the same measure of body composition be used over time. (C) 2018 by the National Kidney Foundation, Inc. All rights reserved.