Assessment of bioelectrical impedance analysis devices for data reliability of body impedance measurements

Background: Bioelectrical impedance analysis (BIA) devices are widely used for body composition analyses, but their accuracy for body impedance measurements can vary significantly between BIA device models. Calibration of BIA device is essential to ensure the reliability of measurement data, particularly for clinical and research purposes. This study aims to evaluate the electric impedance measurement abilities of multi-frequency BIA devices using a home-built body impedance imitator (BII), designed to simulate human body impedance, and to assess the measurement uncertainty based on calibration with internationally accredited reference standards.

Methods: Two different models of multi-frequency BIA device with similar specifications and measurement schemes were evaluated using the BII as a reference standard. First, the impedance values of the BII were calibrated using an LCR meter, traceable to national impedance standards. Next, the impedance measurements were performed using both models based on the standard BII over a wide frequency range (1 kHz to 1 MHz) to analyze the accuracy, linearity, contact resistance effects, etc. of the BIA devices.

Results: Despite having similar specifications, two models of BIA device exhibited notable differences in their measurement performance. Model A showed better accuracy and consistency, with measurement uncertainty of approximately 1.4% at a 95% confidence level (k = 2). In contrast, Model B showed higher deviations, particularly in trunk impedance measurements. The study found that the BIA device calibration procedure could effectively quantify the measurement uncertainty and ensure the reliability of measurement data provided by the BIA device.

Conclusion: This study demonstrates that BIA devices can be calibrated using a standard procedure with a stable reference such as the BII, which traces back to internationally accredited impedance standards. The results show that the calibration significantly improves the reliability of BIA device measurements, providing a level of confidence for the measurement data.

1 Introduction

So called bioelectrical impedance analysis (BIA) devices are now widely used in hospitals for medical checkups and clinical examinations, as well as in fitness centers or even at home. Since the advent of the BIA, research studies have been actively conducted in terms of instrument developments and clinical applications (1–14). As a result, the BIA technology has matured over last several decades, and its industry has expanded significantly along with the fast growing health industry.

Since a BIA device estimates the body composition, such as body cell mass, total body water, and fat-free body mass, etc., based on the electrically measured body impedance, its impedance measurement capability is fundamental in determining the instrument’s basic specifications. Thus, when assessing the accuracy of a BIA device, its impedance measurement performance should be primarily evaluated. Furthermore, as the BIA device impedance measurements can be affected by various factors such as electrode contact, environmental conditions, and inherent characteristics of the device itself, an accurate calibration process is required in order to ensure reliability of the body impedance measurement data.

Yet, previous studies on the measurement reliability assessments of BIA devices focused mostly on the final data of body composition analysis, in comparison with the other well-known methods, such as dual energy X-ray absorptiometry (15–21), and less on the raw data of body impedance. Only a few studies have been conducted to explore the accuracy of BIA devices in electrical impedance measurements including the research conducted at National Institute of Standards and Technology from a metrology perspective (22–24).

The BIA technology is increasingly recognized as a valuable clinical tool from the basic nutritional status assessment to clinical applications such as hydration monitoring in dialysis and oncology (25–29). Such expansion can be supported by studies on the reliability of BIA devices. Hence, it is essential to know how accurately a BIA device can measure an individual’s body composition. When the instrument’s measurement accuracy is evaluated based on references, which are traceable to national standards, the measurement data will obtain certified reliability, potentially with an international equivalence. So far, most BIA device manufacturers have paid little attention to such verifications since there are no strict legal requirements. However, the situation is expected to change due to the recent increase in regulatory demands. The BIA device manufacturing industry can move toward adapting the concept of standardization, which provides a reliability in terms of manufacture, safety, and data.

Our interests lie within exploring the confidence level of the electrical impedance data, measured by BIA devices. In advance, we investigated the performance of two models of BIA device on the same human body and found some notable differences in their measured impedance values despite of their similar measurement principles and claimed capabilities. The results will be presented again in detail in the next section. At this point, such distinction between the measurement data highlights the motivation of this study.

Therefore, this study aims to provide the calibration uncertainty of BIA devices in order to confirm that their measurement data can be trusted with a high level of confidence. First, basic principles of the BIA were examined to devise a method for the BIA device assessment. As a result, a body impedance imitator (BII), simulating a human body in terms of electric impedance, was developed to be employed as a reference standard in BIA device calibrations. A BIA device calibration was then conducted using the developed BII, and its calibration uncertainty was analyzed.

2 Methods

2.1 Body impedance measurement principle

A human body reacts to an electric current like an electrical impedance Z, which is a complex quantity, composed of resistance R and reactance X. Various kinds of body tissues respond differently to the flowing current according to their electrical properties. For instance, muscles with a higher water content experience less resistance compared to the relatively water-free tissues such as fat. The reactance, coming from the capacitance of cell membrane increases with the cell number (30). Moreover, the body impedance varies with the frequency of the current flow. At low frequencies below 1 ~ 5 kHz, extracellular fluids primarily determine the impedance. In contrast, as the frequency increases, the cell membrane capacitance decreases, and the current starts flowing through both the extracellular and the intracellular fluids (2, 9, 31, 32).

Based on these characteristics, the body impedance measurements across a wide frequency range enable the body composition analysis. Accordingly, body impedance models and measurement methods have been investigated to achieve optimal performance under ideal conditions. In general, high performance BIA devices measure the electrical impedance of human body using the eight-polar configuration. The basic measurement principle is illustrated in Figure 1A. The body impedance model divides the body into five parts; right arm (RA), left arm (LA), left leg (LL), right leg (RL), and trunk (TR). Each part is connected to two electrodes, either on a hand or a foot, except for the trunk, and the impedance of each part is measured in a 4-terminal way by applying current through two electrodes and measuring voltage across the other two electrodes. For instance, when measuring the impedance of LA (Z_LA), a current is applied between the electrodes of 2 and 5 while voltage is measured between the electrodes of 1 and 3. Impedance of all body parts can be measured in a similar way, one after the other as described in Figure 1B.

Figure 1

Figure 1. (A) Body impedance measurement schematic of an eight-polar bioelectrical impedance analysis device for five body parts—right arm (Z_RA), left arm (Z_LA), left leg (Z_LL), right leg (Z_RL), and trunk (Z_TR)—with eight electrodes at hands and feet. (B) Among the eight electrodes, four are selected in sequence to measure the impedance of each body part using a 4-terminal method. For example, Z_LA can be measured by applying current from electrode 2 to 5 while measuring the voltage across electrode 1 and 3.

In this work, two commercial BIA device models of A and B, both measuring the segmental impedance of five body parts based on the eight-polar configuration at multi-frequencies, were studied. The main objective of the study is to demonstrate the necessity of device calibration through comparative analysis. The BIA devices examined are the ACCUNIQ BC720 and the InBody 770, which are widely used in hospitals and health centers for professional analyses, including clinical trials. These models are produced by different manufacturers, but they basically work based on similar measurement principles and exhibit comparable specifications with measurement ranges of (10 Ω ~ 1,000 Ω) and (1 kHz ~ 1 MHz). Although they each claim similar measurement abilities for the body impedance, they have not been evaluated yet on the same basis of an accredited system. Through our study of device evaluation and calibration, the measurement performance of these two BIA devices can be compared equivalently based on a standard procedure.

Priorly, two models of BIA devices were employed to measure the impedance of the same human body. The measurements were made consecutively with a short time interval in order to have the human body condition as same as possible. Figure 2 presents relative differences between the measured impedance values of the two BIA devices for all five body parts against the frequency. The differences were less than 1% for both arms while they were larger for legs especially at the highest frequency of 1 MHz. The difference in TR impedance was about one order of magnitude higher, reaching up to 22% at maximum as shown in the inset of Figure 2. The larger difference in TR impedance is mainly due to the smaller impedance of TR. While the limbs have impedances in the range of several hundred ohms, the TR impedance is roughly one order of magnitude lower. Thus, 1 Ω difference in the impedance measurement corresponds to a few percentage errors for the limbs but several tens of percentage error for the TR, respectively. Such a comparison demonstrates differences in the impedance measurements between the two BIA devices.

Figure 2

Figure 2. Impedance measurement results for the same human body using two different models of body impedance analysis device. Relative differences between the models are plotted against the frequency for all five body parts. The inset shows the trunk part separately.

Unfortunately, a human body is not an ideal standard reference even if the same human body is used. This is because human body is an active system, which constantly changes by many factors both in short- and long-terms. Hence, it is recommended to introduce a BII as an alternative. Unlike a human body whose impedance varies depending on body hydration level, body posture, body condition, skin moisture level of hands and feet, etc. (33–38), an imitator consisting of passive elements of resistors and capacitors would be more stable and practical in many ways.

2.2 Body impedance imitator development

In purpose of calibrating BIA devices using a reference, a reliable BII is necessary. Since there are currently no commercial BIIs available for sale, studies were carried out to develop one for the BIA device calibration.

Figure 3 shows a simple circuit model for the body impedance. As aforementioned, the body impedance can be divided into five parts of four limbs and one trunk, and each part is modeled as a homogeneous conductor, consisting of a resistor (R_P) in parallel, a series resistor (R_S), and a series capacitor (C_S). R_P represents the extracellular fluid while R_S and C_S represent the intracellular fluid and the cell membrane, respectively (22). At low frequencies below 1 ~ 5 kHz, most current flows through the extracellular path while at high frequencies, current can pass through the cell membrane as its impedance decreases (2, 9, 31, 32).

Figure 3

Figure 3. Basic circuit model of a human body impedance, divided into five parts of right arm (RA), left arm (LA), left leg (LL), right leg (RL), and trunk (TR), each consisting of a resistor (R_P) in parallel, a series resistor (R_S), and a series capacitor (C_S).

2.3 Variable body impedance imitator (VBII)

At the beginning, the imitator was constructed with fixed values for each body part. Impedance of each body part was determined according to actual body impedance of an adult male. After conducting some initial tests, the imitator was upgraded to include a variable feature, allowing for more convenient evaluation of BIA devices over a wide impedance range as real human body exhibits different spectrum of body impedance according to the sex, age, pathological condition, etc. The variable body impedance imitator (VBII) was constructed with manually adjustable five channels, corresponding to five body parts. Each body part consists of R_P, R_S, and C_S. The dynamic range of each channel was designed to be from 1 Ω to 1,000 Ω in resistance and from 1 pF to 1 μF in capacitance. Figure 4A shows impedance changes of the VBII against the frequency. The limb impedances were measured using an LCR (Inductance L, Capacitance C and Resistance R) meter and the BIA device model A, and they changed more than 35% at the highest frequency of 1 MHz when compared to the lowest frequency of 1 kHz as can be seen more closely in the inset. Although the impedance values measured by the BIA device model A matched well with those measured by the LCR meter, the frequency dependency of VBII was rather large. Since a real body impedance does not exhibit such a large frequency dependency, it was decided to further improve the imitator.

Figure 4

Figure 4. Relative changes of impedance values, measured by a body impedance analysis device, against the frequency for different body impedance imitators of (A) variable body impedance imitator and (B) programmable body impedance imitator.

2.4 Programmable body impedance imitator (PBII)

An improved programmable body impedance imitator (PBII) was designed to have a negligible frequency dependency within the measurement range utilizing the series RLC resonance effect. In a series RLC circuit, the inductive reactance of the inductor becomes equal to the capacitive reactance of the capacitor (X_L = X_C) at a specific frequency point. This results in the cancellation of the two reactances, and the impedance of the series RLC circuit becomes purely “real” at the resonance frequency. Then, the total impedance of the series RLC circuit corresponds only to the resistance, and the frequency effect becomes negligible.

As presented in Figure 4B, the frequency dependency of the PBII, measured by the BIA device model A, remained flat within 1% over the whole frequency range and for all body parts. In addition, the impedance values of the PBII can be set via a PC with a deviation less than 0.01% (TR: 2%) from the nominal value. The PBII is also designed for on-site calibration, so its physical size is compact for convenient transport.

3 Body impedance analysis device evaluation results

Two aforementioned representative models of BIA devices were evaluated using the developed BIIs in terms of measurement accuracy, resolution, linearity, contact resistance, etc.

3.1 Impedance measurement accuracy

First, impedance values of the BII were measured using both BIA device models of A and B. Although each BIA device measured the same imitator, two models showed distinct frequency dependencies for all body parts as can be seen in Figure 5A. Impedance measurement results in absolute values are also plotted in Figure 5B for the trunk along with LCR meter measurement results. The trunk impedance values, measured by the model A agree better with the reference impedance values, measured by an LCR meter. At 1 kHz, the measured trunk impedance value for the model B differs by about 20% from that of the model A.

Figure 5

Figure 5. Measurement results for the same impedance imitator using two different body impedance analysis devices of model A and B. (A) Relative changes in the measured impedance values with respect to 1 kHz for all five body parts and (B) absolute values of the measured impedance for the trunk against the frequency.

In another measurements using the VBII, differences between two BIA device models in the impedance and reactance values were noticed again as plotted in Figure 6 for all five body parts at 50 kHz. The differences were as large as 20% for the limbs, and they were more than 50% for the trunk, observed similarly in both impedance and reactance. This difference in the impedance measurements also resulted in different body composition analyses; there was about 50% difference in the calculated body water contents between the BIA device models.

Figure 6

Figure 6. Measurement results for the same body impedance imitator using two different models of body impedance analysis devices. Relative differences between the models in (A) impedance and (B) reactance values of all five body parts at 50 kHz.

While measurement differences were observed between the BIA device models, it was repeatedly confirmed that the measured impedance values of the model A agreed well with those of LCR meter. Figure 7 shows that the measured impedance values between the BIA device model A and the LCR meter agreed with each other within ±1% range for all body parts. The study results indicate that BIA devices may have different measurement accuracies despite of their similar specifications, claimed by manufacturers. Indeed, the BIA device calibration can improve reliability of the measurement data.

Figure 7

Figure 7. Relative differences between the LCR meter and the body impedance analysis device of model A in the measured impedance values of a body impedance imitator for all five body parts.

3.2 Impedance measurement resolution

In purpose of assessing the measurement resolution of the BIA devices, the VBII was employed. As aforementioned, each body part of the imitator consists of R_P, R_S, and C_S elements, as illustrated in Figure 3. Their values were adjusted as shown in legends of Figure 8. Figures 8A,B present impedance values of TR, measured by the models of A and B, respectively. In addition to the distinct frequency responses between the models, it is observed that the model A distinguishes the different impedance values of the imitator set as R_P = 20 Ω and 30 Ω, whereas the model B does not. Figures 8C,D present impedance values of LL, measured by the models of A and B, respectively. Here, since the LL impedance is about one order larger than the TR impedance, the model B was able to distinguish the different impedance values set as R_P = 400 Ω and 800 Ω. Yet, the difference measured by the model B was only about 100 Ω at 1 kHz. In case of C_S, the model B appears unable to discriminate a difference of 100 nF at all frequencies while the model A shows some difference at low and high frequencies. Overall, the model A seems to have a better measurement resolution by about one order of magnitude.

Figure 8

Figure 8. Trunk impedance (Z_TR) values measured by different body impedance analysis devices of model (A) A and (B) B against the frequency. Left leg impedance (Z_LL) values measured by different body impedance analysis devices of model (C) A and (D) B against the frequency. Resistor and capacitor elements of a variable body impedance imitator were adjusted to test the measurement resolution as described in the legends.

3.3 Linearity

The VBII was further employed to evaluate the linearity of BIA devices by varying values of the elements of R_P, R_S, and C_S. While the value of one element had been varied, the values of the other two were remain fixed at their values. The measurement results are presented in Figure 9 for the model A. When R_P was varied from 100 Ω to 1,000 Ω, the impedance increased almost linearly as shown in Figure 9A. Variations in R_S and C_S also gave expectable changes in the impedance changes as presented in Figures 9B,C.

Figure 9

Figure 9. Impedance measurement results of the body impedance analysis device model A at 50 kHz by varying (A) parallel resistance (R_P), (B) series resistance (R_S), and (C) series capacitor (C_S) of the variable body impedance imitator.

3.4 Contact resistance

BIA device measurements can be affected by contact resistance between the electrode and the skin of human body being measured, which is why manufacturers emphasize the importance of proper electrode contact in their instrument manuals. Hence, studies were done to explore the contact resistance effects by attaching resistors in series to the electrodes. Various combinations were measured by the BIA device model A, and one representative case is presented here. Series resistors were attached at the potential electrodes of LA and RA (R_POT) and the current electrodes of LA and LL (R_CUR) as illustrated in Figure 10A. The series resistances were varied from 0 to 1,000 Ω, and Figure 10 shows relative changes of (B) TR, (C) LA, and (D) LL impedance values with respect to the zero (no series resistor) series resistor condition against the varying series resistances at 250 kHz. Despite the series resistors, the impedance measurement values for TR changed by less than 1% as shown in Figure 10B and even less for RA and RL. Meanwhile, the measured impedance values of LA and LL were changed up to 5% at certain conditions. It is likely due to the fact that the series resistors were connected to the current electrodes of LA and LL.

Figure 10

Figure 10. (A) Experimental schematic of resistors attached in series to the electrodes of the body impedance analysis device: R_POT: series resistor at the potential electrode (0–1,000 Ω) and R_CUR: series resistor at the current electrode (0–1,000 Ω). Relative changes of (B) trunk (Z_TR), (C) left arm (Z_LA), and (D) left leg (Z_LL) impedance values with respect to the zero series (no series resistor) resistor conditions against the varying series resistances at 250 kHz.

Among the measurement data, a case with 100 Ω resistors attached to both the current and the potential electrodes was selected to further examine the influence of contact resistance over the whole frequency range. Relative deviations in the measured impedance with respect to the zero series resistor condition were plotted against the frequency for Z_LA and Z_LL as shown in Figure 11. The impedance values deviated more as the measurement frequency increased; they decreased by more than 7% at 1 MHz. The larger deviation at the higher frequency might be attributed to the lower impedance contribution of the capacitance component as the capacitive impedance decreases with the increasing frequency.

Figure 11

Figure 11. Relative changes of left arm (Z_LA) and left leg (Z_LL) impedance values with respect to the zero series resistor (no series resistor) conditions against the frequency. Series resistors of 100 Ω were attached at the current electrodes of left arm (LA) and left leg (LL) and at the potential electrodes of left arm (LA) and right arm (RA).

Furthermore, the effects of the contact resistance of the two BIA models of A and B were compared with series resistors at the electrodes. The results are presented in Figure 12, where relative changes in the measured impedance values at 250 kHz are plotted against the series resistance at the current electrodes of LA and LL for different series resistances at the potential electrodes of LA and RA. It was interesting to observe that two models showed rather opposite behaviors. For instance, it can be seen in Figure 12A that the measured Z_LA changed more in the model A when the resistance of the series resistors at the potential electrodes were smaller, while the opposite was observed in model B. Additionally, sign of the impedance change was reversed. On the other hand, for the right body parts of RA and RL, the measured impedance of model A changed slightly, but that of model B changed significantly. It seems like that the correlation in impedance measurements between body parts is relatively high for model B. Again, two BIA models reacted quite differently to the series resistor attachments.

Figure 12

Figure 12. Relative changes of (A) left arm (Z_LA), (B) right arm (Z_RA), (C) left leg (Z_LL), and (D) right leg (Z_RL) impedance values with respect to the zero series resistor (no series resistor) conditions against the varying series resistances at 250 kHz for two different body impedance analysis devices of model A and B. The series resistors were attached at the potential electrodes of LA & RA and the current electrodes of LA & LL as shown in Figure 10A.

In terms of the body composition results, analyzed based on the impedance measurements, the amount of body fat changed by 10% when the measured impedance was changed by 10% for both BIA models though sign of the change was opposite.

3.5 Temperature and humidity

Since the resistors and capacitors of the imitator are passive elements with temperature and humidity coefficients, the environmental effects were also evaluated. Thus, the imitator was placed in a chamber where both temperature and humidity can be controlled, and its impedance was measured using the BIA model A while varying the environmental condition. First, the temperature was varied from 10 °C to 40 °C with the humidity fixed at 50% RH. The impedance remained the same within 2% range for all body parts over the whole measurement frequency range despite the temperature variation. Next, the humidity was varied from 35 to 80% with the temperature fixed at 25 °C. Again, similar results were observed; negligible impedance changes within 1.5% range for the varying humidity levels. Thus, the stability of the BII was evaluated against the temperature and the humidity.

4 Body impedance analysis device calibration and discussions

Based on the previous investigations, a calibration process is proposed for a BIA device using a BII, as illustrated in Figure 13. A BIA device can be calibrated using a BII whose impedance is already calibrated using a reference LCR meter.

Figure 13

Figure 13. Calibration process for a body impedance analysis device using a body impedance imitator. The numbers in the parentheses represent the calibration uncertainty level of each impedance standards or meters.

When calibrating the LCR meter, impedance standards with known reference values (capacitance C, AC resistance AC R, and inductance L) are employed, which are traceable to national resistance standards (AC or DC quantum Hall resistance) (39–41). For instance, calculable resistance standards, whose frequency dependencies can be calculated based on their geometric parameters, can be employed. Well-made calculable standards in the metrology grades exhibit AC values, which deviate only slightly from DC values; AC-DC difference levels of 10⁻⁹ near 1 kHz and 10⁻⁴ at 1 MHz (42).

Typical LCR meters can measure the electrical impedance up to a few MHz, and their impedance measurement range well covers the BIA device measurement range of 10 Ω and 1,000 Ω. Figure 14 shows measurement results for calculable resistance standards of 10 Ω and 100 Ω using an LCR meter. Relative changes in the resistance values with respect to the initial frequency point of 60 Hz are plotted. As seen, some frequency dependency can be observed on level of 0.1% for 100 Ω and 0.5% for 10 Ω, respectively, at the highest frequency of 1 MHz. The inset shows that the change in the 100 Ω resistance stayed within 0.01% up to 100 kHz. It infers that the LCR meter is capable of measuring impedance values with a confidence level of 0.01% at the best. Hence, the developed BIIs were evaluated using the calibrated LCR meter. Through this unbroken chain of calibrations, a BIA device can be assessed based on the BII for its impedance measurement reliability.

Figure 14

Figure 14. Relative changes of AC–DC calculable resistance values, measured by an LCR meter, against the frequency.

Once the BII is calibrated using the reference LCR meter, it is measured by a BIA device to be calibrated. The impedance measurement uncertainty is given as well to present the confidence level of the measured data. There are several factors to be considered for a precise calibration: stabilization of both the BII and the BIA in the measurement environment and proper contact between the BII and the BIA electrodes.

The measurement uncertainty was analyzed as presented in Table 1 for the BIA device model A. There are two types of uncertainty factors. First, the type A uncertainty represents the measurement repeatability, obtained from a statistical variation calculation of the measurement data. On the other hand, the type B uncertainty factors originate from the BII, the BIA device, environment, etc.

Table 1

Table 1. Calibration uncertainty of the body impedance analysis device model A using a body impedance imitator.

Since the BII is employed as a reference standard, its calibration uncertainty and stability contribute to the BIA device calibration uncertainty. The BII was calibrated using the reference LCR meter, and its stability was studied as both short- and long-term stabilities of the BII must be considered. The short-term stability, which can be affected by the environment factors or the measurement system noise, is already considered in the type A uncertainty. Whereas, the long-term stability needs to be taken into account if the BII is not used immediately after a calibration, because its calibration value may change over time.

The other type B uncertainty factors, arising from the BIA device itself are determined by instrument specifications along with measurement environmental factors such as temperature and connecting lead wires. Each uncertainty factor was carefully evaluated as mostly described in the previous section. As shown in Table 1, accuracy, linearity, and temperature are more contributing factors by about one order of magnitude when compared to other factors. After individually evaluating the uncertainty factors, all the factors were combined by a root sum square. As a result, the expanded uncertainty was calculated to be about 1.4% for the limbs and 2.6% for the trunk with a confidence level of 95% (k = 2).

5 Conclusion

Extensive studies have been performed on the multi-frequency BIA devices to explore the reliability of the measurement data. First of all, a BII was developed as a stable reference for the BIA device calibration, because a real human body cannot be a practical solution. The home-built BII was fabricated according to the body circuit model with five body parts of four limbs and one trunk. Each part consists of R_P, R_S, and C_S, representing the extracellular fluid, the intracellular fluid, and the cell membrane, respectively. The impedance of each body part was designed to be adjustable in the range of 1 Ω to 1,000 Ω.

Using the developed BII, two BIA device models of A and B with similar specifications were evaluated in terms of measurement accuracy, resolution, electrode contact, etc. Both models measure the body impedance with eight-polar configuration at multiple frequencies between 1 kHz and 1 MHz. It was found that two BIA models, despite of their similar measurement principles and claimed capabilities, showed some significant differences in their measurement performance. Such distinction between the measurement data of the BIA devices emphasizes the motivation of this study, and the instrument calibration can ensure the confidence level of the measurement data.

Subsequently, a BIA device calibration process was suggested using the developed BII as a reference standard. The impedance values of the BII were calibrated in advance using an LCR meter, which also got calibrated based on impedance standards, traceable to national resistance standards. Afterwards, a BIA device was calibrated using the calibrated BII standard. Consequently, the BIA device measurement capability could be quantitatively evaluated. For instance, the BIA device model A was calibrated with the expanded uncertainty about 1.4% for the limbs and 2.6% for the trunk with a confidence level of 95% (k = 2).

In conclusion, BIA devices can be calibrated using the reference standards like BII for their impedance measurement capabilities, and their measurement data can be presented with a level of confidence, that is internationally accredited through a certified quality system. When manufacturers or users want to get their BIA devices calibrated, they can send their BIA devices or BII standards to accredited calibration laboratories. It would be easier to send the BII standard, and the calibrated BII standard can be used as a reference to evaluate the BIA device. The developed calibration procedure can be generally applied to BIA devices of eight-polar configuration. For other BIA devices of different configuration, it could be still employed after modifying the BII electrode configuration or connection accordingly.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

Author contributions

DK: Data curation, Formal analysis, Investigation, Project administration, Validation, Visualization, Writing – original draft, Writing – review & editing. SS: Formal analysis, Investigation, Writing – review & editing. W-SK: Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Visualization, Writing – review & editing.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This research was funded in part by the Technology Development for Reliability Improvement of Clinical Data under Grant NRF-2018M3A9H6081482 and in part by the Research on Redefinition of SI Base Units under Grant KRISS-2025-GP2025-0001.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

1. Lukaski, HC, Johnson, PE, Bolonchuk, WW, and Lykken, GI. Assessment of fat-free mass using bioelectrical impedance measurements of the human body. Am J Clin Nutr. (1985) 41:810–7. doi: 10.1093/ajcn/41.4.810,

PubMed Abstract | Crossref Full Text | Google Scholar

2. Cornish, BH, Thomas, BJ, and Ward, LC. Improved prediction of extracellular and total body water using impedance loci generated by multiple frequency bioelectrical impedance analysis. Phys Med Biol. (1993) 38:337–46. doi: 10.1088/0031-9155/38/3/001,

PubMed Abstract | Crossref Full Text | Google Scholar

3. Foster, KR, and Lukaski, HC. Whole-body impedance: what does it measure? Am J Clin Nutr. (1996) 64:388S–96S. doi: 10.1093/ajcn/64.3.388S,

PubMed Abstract | Crossref Full Text | Google Scholar

4. Scharfettery, H, Hartingery, P, Hinghofer-Szalkayz, H, and Hutteny, H. A model of artefacts produced by stray capacitance during whole body or segmental bioimpedance spectroscopy. Physiol Meas. (1998) 19:247–61. doi: 10.1088/0967-3334/19/2/012,

PubMed Abstract | Crossref Full Text | Google Scholar

5. Ellis, KJ, Shypailo, RJ, and Wong, WW. Measurement of body water by multifrequency bioelectrical impedance spectroscopy in a multiethnic pediatric population. Am J Clin Nutr. (1990) 70:847–53. doi: 10.1093/ajcn/70.5.847,

PubMed Abstract | Crossref Full Text | Google Scholar

6. Bera, TK. Bioelectrical impedance methods for noninvasive health monitoring: a review. J Med Eng. (2014) 2014:1–28. doi: 10.1155/2014/381251,

PubMed Abstract | Crossref Full Text | Google Scholar

7. Kyle, UG, Bosaeus, I, De Lorenzo, A, Deurenberg, P, Elia, M, Gomez, JM, et al. Composition of the ESPEN working group. Bioelectrical impedance analysis—part I: review of principles and methods. Clin Nutr. (2004) 23:1226–43. doi: 10.1016/j.clnu.2004.06.004,

PubMed Abstract | Crossref Full Text | Google Scholar

8. Guida, B, Laccetti, R, Gerardi, C, Trio, R, Perrino, NR, Strazzullo, P, et al. Bioelectrical impedance analysis and age-related differences of body composition in the elderly. Nutr Metab Cardiovasc Dis. (2007) 17:175–80. doi: 10.1016/j.numecd.2005.11.001,

PubMed Abstract | Crossref Full Text | Google Scholar

9. Matthie, JR. Bioimpedance measurements of human body composition: critical analysis and outlook. Expert Rev Med Devices. (2008) 5:239–61. doi: 10.1586/17434440.5.2.239,

PubMed Abstract | Crossref Full Text | Google Scholar

10. Yeh, C, Chen, YJ, Lai, LY, Jang, TR, Chiang, J, Chen, YY, et al. Bioelectrical impedance analysis in a mathematical model for estimating fat-free mass in multiple segments in elderly Taiwanese males. Int J Gerontol. (2012) 6:273–7. doi: 10.1016/j.ijge.2012.01.031

Crossref Full Text | Google Scholar

11. Chinen, K, Kinjo, I, Zamami, A, Irei, K, and Nagayama, K. New equivalent-electrical circuit model and a practical measurement method for human body impedance. Biomed Mater Eng. (2015) 26:S779-86. doi: 10.3233/BME-151369,

PubMed Abstract | Crossref Full Text | Google Scholar

12. Brantlov, S, Jødal, L, Andersen, RF, Lange, A, Rittig, S, and Ward, LC. Bioimpedance resistance indices and cell membrane capacitance used to assess disease status and cell membrane integrity in children with nephrotic syndrome. Sci World J. (2019) 2019:1–8. doi: 10.1155/2019/4274856,

PubMed Abstract | Crossref Full Text | Google Scholar

13. Nwosu, AC, Mayland, CR, Mason, S, Cox, TF, Varro, A, Stanley, S, et al. Bioelectrical impedance vector analysis (BIVA) as a method to compare body composition differences according to cancer stage and type. Clin Nutr ESPEN. (2019) 30:59–66. doi: 10.1016/j.clnesp.2019.02.006,

PubMed Abstract | Crossref Full Text | Google Scholar

14. Tinsley, GM, Harty, PS, Moore, ML, Grgic, J, Silva, AM, and Sardinha, LB. Changes in total and segmental bioelectrical resistance are correlated with whole-body and segmental changes in lean soft tissue following a resistance training intervention. J Int Soc Sports Nutr. (2019) 16:58. doi: 10.1186/s12970-019-0325-4,

PubMed Abstract | Crossref Full Text | Google Scholar

15. Bracco, D, Thiébaud, D, Chioléro, RL, Landry, M, Burckhardt, P, and Schutz, Y. Segmental body composition assessed by bio-electrical impedance analysis and DEXA in humans. J Appl Physiol. (1996) 81:2580–7. doi: 10.1152/jappl.1996.81.6.2580,

PubMed Abstract | Crossref Full Text | Google Scholar

16. Bedogni, G, Malavolti, M, Severi, S, Poli, M, Mussi, C, Fantuzzi, A, et al. Accuracy of an eight-point tactile-electrode impedance method in the assessment of total body water. Eur J Clin Nutr. (2002) 56:1143–8. doi: 10.1038/sj.ejcn.1601466,

PubMed Abstract | Crossref Full Text | Google Scholar

17. Medici, G, Mussi, C, Fantuzzi, A, Malavolti, M, Albertazzi, A, and Bedogni, G. Accuracy of eight-polar bioelectrical impedance analysis for the assessment of total and appendicular body composition in peritoneal dialysis patients. Eur J Clin Nutr. (2005) 59:932–7. doi: 10.1038/sj.ejcn.1602165,

PubMed Abstract | Crossref Full Text | Google Scholar

18. Bosy-Westphal, A, Later, W, Hitze, B, Sato, T, Kossel, E, Glüer, C, et al. Accuracy of bioelectrical impedance consumer devices for measurement of body composition in comparison to whole body magnetic resonance imaging and dual X-ray absorptiometry. Obes Facts. (2008) 1:319–24. doi: 10.1159/000176061,

PubMed Abstract | Crossref Full Text | Google Scholar

19. Ling, CHY, de Craen, AJM, Slagboom, PE, Gunn, DA, Stokkel, MP, Westendorp, RGJ, et al. Accuracy of direct segmental multi-frequency bioimpedance analysis in the assessment of total body and segmental body composition in middle-aged adult population. Clin Nutr. (2011) 30:610–5. doi: 10.1016/j.clnu.2011.04.001,

PubMed Abstract | Crossref Full Text | Google Scholar

20. Lee, JB, Sung, BJ, Ko, BG, Cho, EH, and Seo, TB. A comparative study on the reliability and validity of body composition results by impedance method measurement device. J Exerc Rehabil. (2023) 19:299–308. doi: 10.12965/jer.2346404.202,

PubMed Abstract | Crossref Full Text | Google Scholar

21. Looney, DP, Schafer, EA, Chapman, CL, Pryor, RR, Potter, AW, Roberts, BM, et al. Reliability, biological variability, and accuracy of multi-frequency bioelectrical impedance analysis for measuring body composition components. Front Nutr. (2024) 11:1491931. doi: 10.3389/fnut.2024.1491931,

PubMed Abstract | Crossref Full Text | Google Scholar

22. Oldham, NM. Overview of bioelectrical impedance analyzers. Am J Clin Nutr. (1996) 64:405S–12S. doi: 10.1093/ajcn/64.3.405S,

PubMed Abstract | Crossref Full Text | Google Scholar

23. Freeborna, TJ, Milligana, A, and Esco, MR. Evaluation of ImpediMed SFB7 BIS device for low-impedance measurements. Measurement. (2018) 129:20–30. doi: 10.1016/j.measurement.2018.07.010

Crossref Full Text | Google Scholar

24. Marcotuli, V, Zago, M, Moorhead, AP, Vespasiani, M, Vespasiani, G, and Tarabini, M. Metrological characterization of instruments for body impedance analysis. Acta IMEKO. (2022) 11:1–7. doi: 10.21014/acta_imeko.v11i3.1179

Crossref Full Text | Google Scholar

25. Son, JW, Han, BD, Bennett, JP, Heymsfield, S, and Lim, S. Development and clinical application of bioelectrical impedance analysis method for body composition assessment. Obes Rev. (2025) 26:e13844. doi: 10.1111/obr.13844,

PubMed Abstract | Crossref Full Text | Google Scholar

26. Branco, MG, Mateus, C, Capelas, ML, Pimenta, N, Santos, T, Mäkitie, A, et al. Bioelectrical impedance analysis (BIA) for the assessment of body composition in oncology: a scoping review. Nutrients. (2023) 15:4792. doi: 10.3390/nu15224792,

PubMed Abstract | Crossref Full Text | Google Scholar

27. Porta, EL, Faragli, A, Herrmann, A, Muzio, FPL, Estienne, L, Nigra, SG, et al. Bioimpedance analysis in CKD and HF patients: a critical review of benefits, limitations, and future directions. J Clin Med. (2024) 21:6502. doi: 10.3390/jcm13216502

Crossref Full Text | Google Scholar

28. Nwosu, AC, Stanley, S, Mayland, CM, Mason, S, McDougall, A, and Ellershaw, JE. Non-invasive technology to assess hydration status in advanced cancer to explore relationships between fluid status and symptoms: an observational study using bioelectrical impedance analysis. BMC Palliat Care. (2024) 23:209. doi: 10.1186/s12904-024-01542-z

Crossref Full Text | Google Scholar

29. Prior-Sánchez, I, Herrera-Martínez, AD, Zarco-Martín, MT, Fernández-Jiménez, R, Gonzalo-Marín, M, Muñoz-Garach, A, et al. Prognostic value of bioelectrical impedance analysis in head and neck cancer patients undergoing radiotherapy: a VALOR® study. Front Nutr. (2024) 11:01–13. doi: 10.3389/fnut.2024.1335052,

PubMed Abstract | Crossref Full Text | Google Scholar

30. Abasi, S, Aggas, JR, Garayar-Leyva, GG, Walther, BK, and Guiseppi-Elie, A. Bioelectrical impedance spectroscopy for monitoring mammalian cells and tissues under different frequency domains: a review. ACS Meas Sci Au. (2022) 2:495–516. doi: 10.1021/acsmeasuresciau.2c00033,

PubMed Abstract | Crossref Full Text | Google Scholar

31. Chumlea, WC, and Guo, SS. Bioelectrical impedance: a history, research issues, and recent consensus. Washington, DC: National Academies Press (US) (1997). 7:169.

Google Scholar

32. Brantlov, S, Ward, LC, Isidor, S, Hvas, CL, Rud, CL, and Jødal, L. Cell membrane capacitance (cm) measured by bioimpedance spectroscopy (BIS): a narrative review of its clinical relevance and biomarker potential. Sensors. (2025) 25:4362. doi: 10.3390/s25144362,

PubMed Abstract | Crossref Full Text | Google Scholar

33. Deurenberg, P, Weststrate, JA, Paymans, I, and van der Kooy, K. Factors affecting bioelectrical impedance measurements in humans. Eur J Clin Nutr. (1988) 42:1017–22. doi: 10.1038/ejcn.1988.136

PubMed Abstract | Crossref Full Text | Google Scholar

34. Roos, AN, Westendorp, RGJ, and Froehlich, M. Tetrapolar body impedance is influenced by body posture and plasma sodium concentration. Eur J Clin Nutr. (1991) 46:53–60. doi: 10.1038/ejcn.1992.8

Crossref Full Text | Google Scholar

35. Liang, MT, and Norris, S. Effects of skin blood flow and temperature on bioelectric impedance after exercise. Med Sci Sports Exerc. (1993) 25:1231–9. doi: 10.1249/00005768-199311000-00005,

PubMed Abstract | Crossref Full Text | Google Scholar

36. Di Iorio, BR, Terracciano, V, and Bellizzi, V. Bioelectrical impedance measurement: errors and artifacts. J Ren Nutr. (1999) 9:192–7. doi: 10.1016/S1051-2276(99)90033-X

Crossref Full Text | Google Scholar

37. Slinde, F, Bark, A, Jansson, J, and Rossander-Hultheln, L. Bioelectrical impedance variation in healthy subjects during 12 h in the supine position. Clin Nutr. (2003) 22:153–7. doi: 10.1054/clnu.2002.0616,

PubMed Abstract | Crossref Full Text | Google Scholar

38. Dixon, CB, Ramos, L, Fitzgerald, E, Reppert, D, and Andreacci, JL. The effect of acute fluid consumption on measures of impedance and percent body fat estimated using segmental bioelectrical impedance analysis. Eur J Clin Nutr. (2009) 63:1115–22. doi: 10.1038/ejcn.2009.42,

PubMed Abstract | Crossref Full Text | Google Scholar

39. Kim, DB, Kassim, DM, Kim, WS, Callegaro, L, D'Elia, V, Trinchera, B, et al. Traceability chain at KRISS from DC quantum hall resistance to farad using coaxial bridges. IEEE Trans Instrum Meas. (2019) 68:1941. doi: 10.1109/TIM.2019.2896365

Crossref Full Text | Google Scholar

40. Tran, NTM, Kucera, J, Kim, WS, and Kim, DB. Calibration of 10 nF capacitance standard from DC quantum hall resistance using a digital impedance bridge. Meas Sci Technol. (2023) 34:075009. doi: 10.1088/1361-6501/acc6e2

Crossref Full Text | Google Scholar

41. Kim, DB, Shin, SS, Kim, WS, and Kucera, J. Realization of inductance scale using digital bridges at low frequencies. IEEE Trans Instrum Meas. (2024) 73:1005207. doi: 10.1109/TIM.2024.3364969

Crossref Full Text | Google Scholar

42. Awan, SA, Kibble, BP, and Schurr, J. Coaxial electrical circuits for interference-free measurements. London: IET Electrical Measurement Series (2011).

Google Scholar