Empirical Evidence for Bio-Impedance as a Predictor of Total Body
Water Volume
Keyon Tehrani

Abstract
Using a data set of healthy and unhealthy subjects with the measured total body water volume using traditional isotope dilution methods, regression models to define an empirical relationship between anthropometric data, bio-impedance, and the predicted total body water volume were evaluated. Regression models were developed using different combinations of parameters as with varying rates of success when evaluated for all patients, healthy patients, and medical patients. The best empirical relationship was a multiple regression model of height, weight, and impedance (R = 0.94)
Background
Determining total body water is important in ensuring the body functions efficiently and predicting overall health [1], [2]. Most traditional methods to determine fluid volume involve determining isotope dilution spaces using tritium or dieuterium and quantifying these traces in physiological fluids using a variety of methods. Oxygen-18 has also been used as a tracer to measure fluid body weight [3]. These methods are costly, timely, and not practical, and a more efficient technique is necessary.
Evaluating the water volume can also be useful in hemodialysis. In hemodialysis, the fluid volume needs to be measured to make sure the procedure is being run accordingly [4]. A fast, and non-invasive technique will be very useful in these clinical settings. Using bio-impedance, the change in fluid volume can be calculated before and after the procedure. Measuring the body

water volume can also be useful in monitoring volume change before and after exercise. Through bio-impedance, water volume can be a monitored on subjects during exercise regimes to evaluate thermoregulation and volume depletion.
Bioelectric impedance analysis can be used to estimate body water compartments for both healthy individuals and those affected by pathologic conditions. A total body water volume 15% or more below the normal value could be an indication of a medical problem
[5].
The theoretical idea that body impedance can be used to predict total body water is based on the concept that the impedance of a simple geometric system is a function of conductor configuration, length, cross-sectional area, and the signal frequency. If the signal frequency is fixed and the conductor configuration is constant, then the impedance is a function of conductor length and cross-section, which is the conductor volume. This is defined by z = ρl/A where z is the impedance, ρ is the resistivity, l is the length, and A is the area. Multiplying by L/L and solving for V yields
V = ρL2/Z.
Using this concept the impedance can be related to the conductor volume [6].
Although the human body is not a simple or uniform geometric conductor, recent studies have been done in using bioelectric impedance to determine the water volume, and have shown that a relationship could exist between impedance and total body volume. value of 0.9487 showing a strong empirical relationship. Approach
The approaches in predicting total body water from electrical impedance and anthropometric parameters involve using several parameters of age, weight, gender, and height combined with the measured bioelectric impedance to develop regression models. These regression models then predict the total water volume.
A data set was obtained from a study conducted at the University of Alabama
Medical Center (Hoffer). This study used the tetrapolar electrical impedance technique, and measured the impedance of subjects in normal health and those with pathological conditions [2]. The total body water of these subjects was measured using the tritium isotope method, and the anthropometric data of height, age, weight, and gender was collected. Using the anthropometric data and the measured bio-impedance, regression tools were used to create different regression models for predicting water volume and computing its accuracy compared to the