Accuracy of Bioimpedance Modalities for Fluid Assessment in Hemodialysis Patients: A Randomized Observational Study

Abstract

Aim Fluid overload is a major contributor to mortality in critically ill patients but is difficult to estimate clinically. Bioimpedance has been used to estimate fluid volumes with three different methods of analysis:1. single-frequency; 2. multi-frequency; 3. bioimpedance spectroscopy. The aim of this study is to assess the accuracy of different types of bioimpedance analysis in detecting changes in fluid volumes. Methods Prospective observational study, in end-stage renal disease patients requiring dialysis, in a tertiary care center. During hemodialysis, we assessed the correlation between change in estimated total body water volumes, as measured by all three methods of bioimpedance, and fluid volumes removed, as measured by changes in body weight. Results Twenty-four pediatric and adult patients were included in the study (median age 42.4 years) with a total of 30 study assessments performed. There was a weak correlation between change in body weight and change in estimated total body water volumes (R=0.15, 0.41, and 0.38, respectively). In the Bland-Altman analysis, the mean biases along with their associated 95% confidence limits of agreement were -0.23 L (-4.1 to 3.5 L) for single-frequency; -1.1 L (-4.1 to 1.9 L) for multi-frequency; and -0.6 L (-6.1 to 4.8 L) for bioimpedance spectroscopy. Conclusion In this study of end-stage renal disease patients requiring dialysis, the accuracy of bioimpedance measurement to evaluate fluid changes was poor, regardless of bioimpedance modality.