National Health and Nutrition Examination Survey

Bioelectrical impedance analysis (BIA) measures the electrical impedance of body tissues and can be used to assess fluid volumes, total body water, and fat-free body mass. The NHANES bio-impedance spectroscopy (BIS) multi-frequency measurements were collected in the BIA examination, together with dual energy x-ray absorptiometry (DXA) scans, as part of the Body Composition component in the Mobile Examination Center (MEC). A small alternating current was passed through surface electrodes placed on the right hand and foot and the impedance to the current flow was measured by different electrodes placed adjacent to the injection electrodes. The voltage drop between electrodes provided a measure of impedance, or opposition to the flow of the electric current.

The BIA examination was administered to eligible survey participants 8-49 years age. Pregnant females were not eligible for the BIA examination. Other reasons for exclusion from the BIA examination were:

• Amputations other than fingers or toes
• Artificial joints, pins, plates, or other types of metal objects in the body
• Pacemaker or automatic defibrillator
• Coronary stents or metal suture material in the heart
• Weight over 300 pounds (limitation for examination table)

Pregnancy was determined by a positive urine pregnancy test and/or self-report at the time of the BIA examination. Females between the ages of 12–49 years and menstruating 8–11 year olds were not permitted to take the BIA examination without a negative MEC pregnancy test result. In addition, females aged 12–49 years were excluded from the examination if they said they were pregnant at the time of the exam, even if the pregnancy test was negative.

The NHANES BIS data were collected with a HYDRA ECF/ICF Bio-Impedance Spectrum Analyzer (Model 4200) manufactured by Xitron Technologies, Inc., San Diego, California. The multi-frequency analyzer uses a full 12-bit digital signal processing technique to measure impedance at 50 frequencies logarithmically spaced from 5KHz to 1 MHz (HYDRA ECF/ICF Bio-Impedance Spectrum Analyzer (Model 4200) Operating Manual).

In the Body Composition component, the DXA examination was conducted first, followed by BIA. Both examinations were conducted by trained health technicians. The examination protocol is documented in the Body Composition Procedures Manual on the NHANES website. The collection of raw frequency data (resistance and reactance) took approximately 15 seconds.

A 6-hour fast is recommended by the manufacturer before measurements are performed (HYDRA ECF/ICF Bio-Impedance Spectrum Analyzer (Model 4200) Operating Manual). At the time of the BIA examination, survey participants could have consumed a recent meal, been fasting for a blood draw in the MEC, or eaten a light snack in the MEC. The variable PHAFSTHR in the laboratory dataset, Phlebotomy Fasting Questions, provides the time in hours between the survey participant last eating or drinking anything other than water and the time of the blood draw. After the blood draw, the survey participant was offered a small carton of juice and a small box of cookies or crackers while completing other examinations such as BIA. If possible, the blood draw was the first examination performed.

Each time the Hydra analyzer was turned on, it performed an automatic internal circuitry self-check that indicated if the device was out of calibration or had a fault. An electronic verification module was used to complete weekly circuit testing. The Body Composition Procedures Manual details the BIA quality control tests performed.

Experienced trainers, National Center for Health Statistics staff, and contractor staff monitored health technician performance in the field. Retraining sessions were conducted with the technicians periodically, as needed, and annually to reinforce the proper protocols and technique.

The quality of the collected raw frequency data was evaluated through an external Hydra modeling program provided by Xitron Technologies, Inc. (HYDRA ECF/ICF Bio-Impedance Spectrum Analyzer (Model 4200) Operating Manual). The program measured resistance and reactance and calculated the reciprocal impedance (Ζ) and phase angle (θ) at each measured frequency.

The Ζ and θ spectra data were fit to a Cole model (Cole) using iterative non-linear curve-fitting software (De Lorenzo et al.). The modeling program evaluated the weighted least square error of both Ζ and θ, where the weighting was established by the published accuracy specifications of the machine, and removed any frequency that significantly decreased the total weighted least square error. The expected accuracy at each measured frequency was established as a pair of arrays (Ζ and θ) at the start of the modeling function. The error ratio was then established at each modeling point by dividing the modeling error by the corresponding stored array errors. In addition to the correlation of fit using scalar Z, the program established the accuracy of data fit to the model. Measured height (BMXHT) and weight (BMXWT) from the NHANES Body Measures component were input into the modeling program. The program computed fluid volume once the raw frequency data were modeled and the ECF and ICF resistance obtained. 
 

The Hydra modeling program can have difficulty solving the Cole model when the subject has very little cell membrane capacitance or a very high characteristic frequency (F_c) (HYDRA ECF/ICF Bio-Impedance Spectrum Analyzer (Model 4200) Operating Manual). The characteristic frequency of a tissue varies depending on its physical structure. For example, extremely high F_c (>1 MHz) have been reported in severly dehydrated, malnourished infants (Mazariegos et al.). With an observed F_c near or greater than the limit of measured high frequency data, it is very difficult to accurately compute ICF, although measurement of ECF does not appear to be affected (Finn et al.).

 The modeling program was not used to conduct data analyses or to calculate estimates, but to determine the quality of the data as one step in the quality assurance process. The BIDFIT variable provides the following codes for the accuracy of the fit to the Cole model and data quality:

0 = Excellent fit
1 = Good fit
2 = Marginal fit
3 = Questionable fit
4 = Bad fit
5 = Was not able to be fit
7 = No BIA data available

Volume data were not computed by the modeling program if the fit was rated as “Questionable,” “Bad,” or not able to be fit.

Before conducting analyses, users are advised to review the BIDFIT values to identify the quality of the data to be analyzed.

The Hydra equations used to predict ECF and ICF from the modeled extracellular and intracellular resistance used Hanai dielectric mixture theory (Hanai) and have been validated numerous times (Van Loan, Withers et al., Patel et al., Plank et al.). However, the validation research was conducted on specific populations with well-designed protocols and under controlled conditions (HYDRA ECF/ICF Bio-Impedance Spectrum Analyzer (Model 4200) Operating Manual). The NHANES data user should be aware that the validation research is applicable only to those subject populations and may not be applicable to other subject groups.

Although the technology used by the Xitron Hydra has been used to accurately predict fat-free mass (Van Loan, Kopp et al., Van Loan), users should be aware that the prediction of fat-free mass with impedance is a second order estimation subject to error and is not strongly advised (HYDRA ECF/ICF Bio-Impedance Spectrum Analyzer (Model 4200) Operating Manual). To predict fat-free mass (FFM) by impedance, total body water is assumed to be 73.2%, although there can be variation even among young healthy subjects. Fat-free mass is predicted using the equation (d_ECF V_ECF) + (d_ICF X V_ICF) where d_ECF and d_ICF are the mean densities of the ECF and ICF and their associated materials. Fat mass (FM) was calculated by subtracting FFM by impedance from the participants’ measured weight. Percent fat was calculated by dividing the participants’ FM by their measured weight. Estimation of percent fat is third order and subject to additional error.

Data were deleted for participants whose FFM values were close to or greater than their measured weight, which resulted in negative FM values or extremely low FM values that were not phyisologically possible. Other FM values considered to be very low, however, were retained in the data file. The user should review the FM values in the data file before conducting analyses.

The variable BIAEXSTS indicates examination status.

1 = Examination completed
2 = Examination incomplete
3 = Not examined, pregnant
4 = Not examined, weight > 300 lbs
5 = Not examined, jewelry, other objects not removed or metal in body
6 = Not examined, other reason

The “Not examined, other reason” code includes no time to complete the examination, pregnancy test not completed, and participant refusal, as well as exclusions for reasons other than those listed.

Please refer to the Analytic Guidelines for further details on the use of sample weights and other analytic issues. The Analytic Guidelines are available on the NHANES website.