Bioelectrical impedance analysis (BIA) is a frequently used method for estimating body composition based on a 2-component model (2C). It measures “impedance”, the opposition to a small electrical current as it travels through the body’s water pool. Impedance comprises both resistance and reactance:
The small electrical current is passed through the body from conductive surfaces or electrodes. Conductivity is higher through fat free mass (which includes muscle, bone and water) than through fat mass which contains very little water. Different body components have varying levels of impedance in response to different frequencies of the electrical current.
Output is commonly provided in the form of an impedance value (expressed in the unit Ohms, Ω; approximate range between 150Ω - 900Ω). Interpretation of the impedance value varies by BIA instrument type. For single frequency BIA, the impedance value is interpreted as resistance (R). For multi-frequency BIA and bioelectrical impedance spectroscopy, two values are provided, one for resistance (R), and one for reactance (Xc).
The impedance value is related to the volume of a conductor (the overall body size) and the square of the length of the conductor (a distance which is a function of the height of the participant). The volume of the body can be estimated from the ratio of its length/height squared divided by the resistance (H2/R, also known as the impedance index), where H is the height of the participant and R the resistance.
The impedance index is related to the volume of body water and typically used to estimate fat free mass (FFM), assuming that 73% of the body’s FFM is water. BIA is used to assess the dimensions shown in Table 1, depending upon the type of instrumentation used. The types of BIA instrument are described in more detail in the section below.
Table 1 Anthropometric dimensions which can be assessed according to BIA instrument type.
|Total fat free mass (FFM)
|Total fat mass (FM)
|Total body water (TBW)
|Extracellular water (ECW)
|Intracellular water (ICW)
|Body cell mass (BMC)
|Central and peripheral FFM and FM
|Total abdominal fat and visceral fat
* Multi-segmental approach only
Bioelectrical impedance analysis (BIA) instruments use contact electrodes that send the electrical signal through the body. These electrodes are either patch types (similar to ECG electrodes) or stainless steel plates. BIA instruments can be broadly classified into three types: single-frequency BIA (SF-BIA); multi-frequency BIA (MF-BIA); bioelectrical impedance spectroscopy (BIS). Some BIA systems are incorporated into digital electronic scales, simultaneously measuring impedance and body weight with a force sensor.
Single-frequency BIA (SF-BIA)
SF-BIA (frequency of 50 kHz) also known as tetrapolar impedance is the most commonly used BIA instrument, based on 4 contact electrodes (2 injecting and 2 sensing electrodes). The impedance is then used together with other anthropometric data, age and gender to predict body composition variables using empirical linear regression equations.
SF-BIA is unable to distinguish the distribution of total body water into its intracellular and extracellular components.
Multi-frequency BIA (MF-BIA)
MF-BIA (frequencies up to 800 kHz) allows differentiation of intracellular and extracellular components of total body water. It relies on the principle that the body's impedance is dependent on the frequency of the alternating current applied. Total body water is distributed between the intracellular and extracellular components, separated by cell membranes. Cell membranes act as capacitors that insulate the intracellular water (ICW) at low frequencies so that predominantly extracellular water (ECW) is measured. At higher frequencies, the membranes are permeable to the current, so that ICW and ECW are both determined. MF-BIA like SF-BIA also uses regression models to evaluate FFM, TBW, ICW and ECW.
ViScan is a tetrapolar impedance method. The abdominal body composition values (total abdominal adiposity and visceral fat) are derived from extrapolation of impedance measures (at 6.25 and 50 KHz) using inbuilt algorithms and waist circumference. It consists of a wireless measurement belt and an infrared beam projected over the waist at the umbilical sagittal plane. It detects the waist circumference using two infrared sensors on either side of the base unit.
Bioelectrical impedance spectroscopy (BIS)
BIS uses a series of frequencies and it is based on the Cole–Cole plot and Hanai models which characterise the measurement segment with parallel circuits for ECW and ICW, and accounts for a capacitive effect introduced by the non-conducting membrane that separates the ICW and ECW. BIS firstly determines the electrical resistance of ECW and ICW, and then calculates the volumes of these respective components. By differentiating between extracellular water and intracellular water spaces, BIS can provide an estimate of body cell mass.
The multi-segmental approach which is based on 8 contact electrodes (2 on each hand and foot) assumes that the body is made up of a group of cylinders (left and right arms, the left and right legs, and the total body are measured) and provides body composition values for the trunk and limbs as well as the whole body. Multi-segmental BIA (SEG-BIA) is available in both single-frequency and multi-frequency body composition monitors.
Figure 1 Electrodes placement on hand and foot.
Source: MRC Epidemiology Unit.
Figure 2 Participant’s feet touching the electrodes on a foot to foot body composition monitor.
Source: MRC Epidemiology Unit.
Figure 3 BIA using the hand to foot 8-electrode body composition monitor.
Source: MRC Epidemiology Unit.
Figure 4 Estimating abdominal fat using ViScan.
Source: MRC Epidemiology Unit.
Reliable BIA requires protocol standardisation and control from the following:
A selection of published BIA equations for predicting FFM, FM, TBW, and ECW, which include description of BIA instruments, the criterion used to validate the equations and standard error of the estimates was published by Kyle et al. (2004).
When deriving fat mass from those equations, the absolute error of estimates at individual level will vary, but the ranking of individuals (i.e. relative validity) will be relatively stable regardless of the equation used.
If, in the analysis, the investigators are only interested in determining ranking of body composition traits rather than absolute values, an alternative approach, which avoids the need for population specific validation equations can be used, as described by Vanltalie et al (1990), Tyrrell et al. (2001) and Wells et al. (2007). This method involves correcting lean and fat masses (kg) for height (cm) by deriving the following indices.
This method provides good rankings for both outcomes.
BIA data are usually expressed in the form HT2/R and then used to predict total body water. Total body water then requires adjustment for the hydration of lean tissue (HLT) to calculate lean mass. The adjustment for HLT assumes a constant level of hydration between individuals.
When using BIA data, the index for fat free mass/lean mass can be expressed as: Lean mass index = 1/R(Ω)
As lean mass is equal to total body water/hydration of lean tissue (HLT) and that total body water is proportional to H2/R, the relationship between R and lean mass index can be summarized as: Lean mass index (H2/R)/ HLT/ H2 = 1/(R*HLT)
If study population is of homogeneous age, HLT is left out as a sex specific constant, resulting in lean mass index being proportional to 1/R. Based on this theoretical approach, individuals of the same sex can be ranked according to this simple BIA index.
Further information in Wells et al. (2007).
An overview of the characteristics of BIA is outlined in Table 2.
Table 2 Characteristics of bioelectric impedance analysis.
|Number of participants
|Researcher burden of data collection
|Researcher burden of coding and data analysis
|Risk of reactivity bias
|Risk of recall bias
|Risk of social desirability bias
|Risk of observer bias
|Suitability for field use
|Participant literacy required
Considerations relating to the use of BIA in specific populations are described in Table 3. Estimates of body composition values are dependent on the validity of the BIA equation used for the population.
Table 3 Anthropometry by BIA in different populations.
|BIA may not be suitable to estimate fat-free mass FFM and fat mass FM due to the hydration status throughout pregnancy. BIA predictions will be limited in their ability to account for this variation. However, it has been used to monitor TBW changes. Some manufactures recommend not using their devices during pregnancy.
|Infancy and lactation
|There is a large variation in the different body components (water, protein, minerals) from birth to adulthood due to growth and biological maturation. This variation can significantly affect the estimate of FM and FFM, in two-compartment models like the BIA method. BIA predictions will be limited to account for these variations. Lack of valid regression equations to predict body fat make this method not suitable for these populations. However, it has been used to monitor TBW changes. Lack of standardisation of electrode placement in infants is also an issue.
|Toddlers and young children
|There is large variation in the different body components (water, protein, minerals) from birth to adulthood due to growth and biological maturation. This variation can significantly affect the estimate of FM and FFM, in two-compartment models like the BIA method. BIA predictions will be limited to account for these variations. There is lack of standardisation of electrode placement in studies. Lack of valid regression equations to predict body fat make this method not suitable for these populations Standing monitors are limited in their use in these populations as the metal electrodes are too large for children’s feet. Many of those monitors are not recommended in children below 7 years of age.
|Suitable, but presence of oedema may affect estimates.
|Suitable but tend to overestimate fatness in lean individuals.
|Suitable but tendency to underestimate fatness in those individuals.
Refer to section: Practical considerations for objective anthropometry