Athletic Training and Sports Health Care

Original Research 

Comparison of Lower Extremity Lean Mass Between Multi-frequency Bioelectrical Impedance Analysis and Dual-Energy X-ray Absorptiometry in Athletic College-Aged Men and Women

Melissa M. Montgomery, PhD, ATC; Amanda J. Tritsch, PhD, ATC, CSCS

Abstract

Purpose:

To compare lower extremity lean mass (LELM) and lean body mass (LBM) between bioelectrical impedance analysis (BIA) devices and dual-energy x-ray absorptiometry (DXA).

Methods:

Thirty-one women and 26 men underwent body composition testing. Inter-instrument reliability and bias were assessed with intraclass correlation coefficients and t tests, respectively.

Results:

Systematic underestimation bias for LELM was observed in men (approximately 3 kg) and women (approximately 1.5 kg) (both P < .01). Wide limits of agreement (LOA) were evident in men (approximately 4.2 kg) and women (approximately 4.9 kg). Overestimation error in men increased with the amount of LELM. Overestimation (P < .01) and proportional (P = .02) biases were noted for LBM in women. Wide LOAs were noted for women (approximately 8 to 9.7 kg) and men (6.7 to 12 kg).

Conclusions:

BIA did not agree well with DXA for assessing LELM, but was more accurate for LBM. BIA consistently underestimated LELM and error depended on the amount of LELM. BIA should be used and interpreted conservatively.

[Athletic Training & Sports Health Care. 20XX;XX:XX–XX.]

Abstract

Purpose:

To compare lower extremity lean mass (LELM) and lean body mass (LBM) between bioelectrical impedance analysis (BIA) devices and dual-energy x-ray absorptiometry (DXA).

Methods:

Thirty-one women and 26 men underwent body composition testing. Inter-instrument reliability and bias were assessed with intraclass correlation coefficients and t tests, respectively.

Results:

Systematic underestimation bias for LELM was observed in men (approximately 3 kg) and women (approximately 1.5 kg) (both P < .01). Wide limits of agreement (LOA) were evident in men (approximately 4.2 kg) and women (approximately 4.9 kg). Overestimation error in men increased with the amount of LELM. Overestimation (P < .01) and proportional (P = .02) biases were noted for LBM in women. Wide LOAs were noted for women (approximately 8 to 9.7 kg) and men (6.7 to 12 kg).

Conclusions:

BIA did not agree well with DXA for assessing LELM, but was more accurate for LBM. BIA consistently underestimated LELM and error depended on the amount of LELM. BIA should be used and interpreted conservatively.

[Athletic Training & Sports Health Care. 20XX;XX:XX–XX.]

Body composition assessment has long been an important component of sports performance and sports medicine practice, with important implications for health and performance in athletes1 and injury risk.2–5 As such, practitioners are consistently interested in the most accurate, affordable, and non-invasive instruments. Although underwater weighing and dual-energy x-ray absorptiometry (DXA) are accepted as the clinical gold standards for total6 and segmental7,8 body composition assessment, respectively, these instruments are often unavailable due to the inherent difficulty in the assessment methods, cost, and operator expertise needed to conduct the assessment. Accordingly, bioelectrical impedance analysis (BIA) has become an increasingly popular technology that allows practitioners to overcome these barriers.

Several studies have investigated the reliability and validity of various BIA instruments; however, because the technology rapidly improves, validation work is continuously needed. For example, multiple-frequency BIA instruments have shown superior agreement with DXA when compared to earlier generation single-frequency technologies.1 Although the cost and portability afforded by BIA appear to be its main benefits, it can also be used for segmental lean mass assessments, which were previously only performed with DXA. This assessment is particularly important given the evidence that muscle mass in the involved limb decreases following anterior cruciate ligament reconstruction,9,10 which may help explain persistent asymmetries in strength and function. Therefore, the ability to incorporate lower extremity lean mass (LELM) assessments into screening and rehabilitation evaluations may yield important information about injury risk and provide opportunities for injury prevention.

Previous investigations of BIA devices have reported acceptable reliability and agreement in LELM in middle-aged11 and older adults,12 but there is limited research on a young, active population. To our knowledge, there are only a few studies that have compared BIA and DXA in this population.13–15 Although one of the studies found small systematic errors in LELM measurements by BIA13 and the other did not,15 they both concluded that BIA was appropriate for assessing LELM. This apparent discrepancy between findings and conclusions highlights a challenge in the literature examining inter-instrument agreement: the judgment is subjective because the investigators must make their own decisions as to whether the error they observe is acceptable, considering the clinical population and setting. Because varying types of error (eg, systematic and proportional bias) are prevalent among BIA devices and appear to be specific to particular populations,1,13–17 new devices should be investigated for validity and reliability when they come on the market to determine whether they are fit to replace DXA, particularly when considering using those assessments to determine injury risk and make important clinical decisions such as an athlete's readiness to return to sport after rehabilitation. Therefore, the primary purpose of this study was to assess the reliability and agreement in LELM between BIA and DXA in healthy, active participants. Due to the minimal published validation work on total body composition assessments in this population with these devices, we also analyzed total lean body mass (LBM).

Methods

Men and women who were physically active (participated in physical activity more than three times per week for more than 30 minutes) were recruited for this study. Participants were required to be between 18 and 30 years of age, euhydrated, and refrained from exercise and food for 4 hours prior to testing. Informed consent was obtained per university institutional review board protocol. Because DXA is contraindicated during pregnancy, women were asked to verbally confirm that they were not pregnant.

When participants arrived at the laboratory, their urine specific gravity was measured with a digital refractometer (Atago USA, Inc., Bellevue, WA) to ensure acceptable hydration status (< 1.025), which is known to influence body composition devices. After confirming acceptable hydration level, participants were asked to empty their bladders. Their height was measured with a wall-mounted stadiometer and the measurement was used in all subsequent body composition tests. Participants wore compression shorts and sports bras (for women) during testing.

Body Composition Testing

Participants completed three body composition tests: two multi-frequency BIA (“770” and “S10” models; In-Body US, Cerritos, CA) and Lunar Prodigy DXA (GE Healthcare, Madison, WI). We chose these particular BIA devices because the BIA 770 is the manufacturer's “research grade” instrument, with the greatest capabilities including fat mass, fat-free mass, and intra-cellular and extra-cellular water. The BIA S10 has the same measurement capabilities, but it uses cabled electrodes that provide enhanced portability, which allows field-based assessments that have been limited in the past with bulkier devices. Both devices take 30 impedance measurements over 6 frequencies via tetrapolar 8-point tactile electrode configurations. Because of the known sensitivity of BIA to fluid pooled in the distal extremities,18 participants completed the BIA 770 test first because they were already in a standing position from their walk to the laboratory and during intake procedures. DXA was performed second and was followed by the BIA S10 because both tests were performed with the participant in a supine position.

Before the BIA 770 test, participants wiped the palms of their hands and the soles of their feet with an antibacterial solution (InBody Tissue; InBody US) to enhance conductivity with the electrodes. Participants stepped on the scale and stood still while their body mass was measured. Their age and height were input into the device. Then participants aligned their heels and forefeet on the electrodes on the measurement scale and their hands and fingers on electrodes on the device handles. They were instructed to extend their elbows and slightly abduct their shoulders so that their arms did not touch their trunk. Participants were asked to stand still and remain silent while the device completed the 30-second measurement.

For DXA, participants laid in a standard supine position recommended by the manufacturer for a total body scan. Height and body mass (obtained from the BIA 770 device) were entered in the EnCORE 2012 software (GE Healthcare) for the body composition calculations. After proper positioning, participants were instructed to lie still for the 6-minute total body scan.

After the DXA test, participants remained in the supine position for the BIA S10 test. The investigator cleaned the participants' fingers and ankles with antibacterial wipes before placing bilateral electrodes posteriorly and inferiorly to the malleoli and the thumb and index finger. The participants' demographics were input into the device. Before beginning the test, the investigator confirmed that participants had been supine for at least 10 minutes. Participants were instructed to lie still and remain silent for the 90-second test.

Data Reduction and Analysis

Lean mass calculated by the BIA devices for LELM and LBM were used for analysis. Because DXA uses a three-compartment model, the sum of the lean mass and bone mineral content from the software was used. To assess inter-instrument reliability, intraclass correlation (ICC2,1)19 and standard error of measurement (SEM) were calculated.20 Reliability was interpreted as “moderate,” “good,” or “excellent” if the ICC value was between 0.5 and 0.75, 0.75 and 0.9, or greater than 0.90, respectively.21

Fixed bias was assessed by calculating the differences in lean mass acquired by BIA and DXA. The mean and the difference between each pair were calculated and confirmed to conform to a normal distribution. Bland– Altman plots22 were created by plotting the mean on the x-axis against the difference on the y-axis. The 95% limits of agreement (LOA) were calculated (±1.96 × SD) and displayed on each plot as a metric of precision with the bias. The plots were visually inspected to assess the relationship between the difference and the mean.

After sample t tests were used to determine whether a significant bias (statistically not equal to zero) was present, linear regressions were used to evaluate proportional bias. A significant regression coefficient identified proportional bias, indicating that the difference between instruments depended on the amount of lean mass. Because of the known significant differences in body composition profiles between men and women23 and previous reports of proportional bias that are dependent on body composition profile,14,17 analyses were performed separately by sex. A significance level of less than .05 was determined a priori and analyses were performed with IBM SPSS software (version 25; IBM Corporation, Armonk, NY).

Results

Thirty-one women (age: 23.0 ± 3.0 years; height: 1.67 ± 0.9 m; weight: 62.8 ± 8.2 kg; body mass index: 22.8 ± 3.0 kg/m2) and 26 men (age: 24.2 ± 3.0 years; height: 1.76 ± 0.9 m; weight: 77.0 ± 9.7 kg; body mass index: 25.1 ± 4.3 kg/m2) completed testing. Although participation was open for all volunteers who met the inclusion criteria, most of the participants were Division I soccer players from our university. Table 1 lists the descriptives for LELM and LBM.

Mean ± SD for LELM and LBM from BIA Devices and DXA

Table 1:

Mean ± SD for LELM and LBM from BIA Devices and DXA

Reliability

ICC (ICC2,1) and SEM are provided in Table 2. For LELM, good reliability was observed in women, whereas moderate reliability was observed in men. The SEM was within 1 kg of lean mass in women, whereas it exceeded 1.2 kg in men. For LBM, both BIA instruments demonstrated excellent reliability and similar precision.

ICC and SEM (kg) for Agreement of LELM and LBM Results Between BIA Devices and DXA

Table 2:

ICC and SEM (kg) for Agreement of LELM and LBM Results Between BIA Devices and DXA

Fixed Bias

Bias and 95% LOA for LELM and LBM are displayed in Tables 34, respectively. A significant fixed bias was observed for both BIA devices compared to DXA. The BIA 770 and BIA S10 devices underestimated LELM in women (t(30) = 7.0; P < .01 and t(30) = 6.0; P < .01, respectively) and men (t(25) = 15.2; P < .01 and t(25) = 13.4; P < .01, respectively). Both devices had similarly wide LOAs that were evident in men and women, spanning more than 4 kg of LELM. A significant fixed bias was also observed for the BIA 770 in women only, which overestimated LBM by 2.34 ± 2.0 kg (t(30) = −6.4; P < .01) on average, whereas the S10 overestimated by 1.54 ± 2.0 kg (t(30) = −4.2; P < .01). No bias was noted for either device in men (P value range: .10 to .92).

Bias and 95% LOA for LELM Compared to DXAa,b

Table 3:

Bias and 95% LOA for LELM Compared to DXA,

Bias and 95% LOA for LBM Compared to DXAa,b

Table 4:

Bias and 95% LOA for LBM Compared to DXA,

Proportional Bias

When analyzing LELM with BIA devices (Table 3), the mean was a significant predictor of the bias in men when analyzing the BIA 770 (F1,25 = 9.3; P < .01) and the BIA S10 (F1,25 = 5.7; P = .03). The positive regression coefficients for men indicated that LELM was overestimated more so in those with greater magnitudes of LELM. This proportional bias was not observed in women (P = .27 to .35).

For LBM measured with the BIA 770, proportional bias was observed in women (F1,30 = 6.3; P = .02) but not men (F1,24 = 0.83; P = .37). Similarly for the BIA S10, LBM predicted the bias in women (F1,29 = 4.9; P =.04), but not men (F1,24 = 0.12; P = .73). Both findings indicate that BIA overestimates more in women with a greater amount of LBM.

Discussion

Our primary finding was that BIA demonstrated moderate-to-good reliability when measuring LELM but significantly underestimated LELM in both men and women. We also observed proportional bias when measuring LELM in men, whereby the underestimation worsened with greater amounts of LELM. We also compared LBM between BIA and DXA and found acceptable agreement; however, there were significant fixed and proportional biases between BIA and DXA in women but not in men. Perhaps the most clinically important finding was the wide LOA between BIA and DXA.

Although comparison with previous work is challenging due to differences in instruments and varying populations, our findings can be primarily discussed relative to the few recent studies in college-aged cohorts.13–15 Because the BIA 770 and S10 devices performed similarly in all analyses when compared to DXA, we discuss both BIA devices together.

LELM

Little research has investigated the agreement in LELM. Moderate-to-good reliability (ICC approximately 0.85) has been reported previously in a middle-aged cohort of men and women11; however, those studies did not report the precision. In the current study, we observed similar reliability, although slightly lower in men. We also noted similar precision in BIA devices in women (approximately 1 kg) and men (approximately 1.5 kg).

Regarding fixed bias (systematic error), we found that BIA devices underestimated LELM by approximately 1.5 kg in women, which is larger than the findings from the two other studies of college-aged women.13,15 In a study of 45 female athletes from varying sports,15 BIA only underestimated the leg lean mass by 0.5 kg, whereas another study reported 0.7 kg. In the current study, we also noted wide LOA (approximately 4.9 kg) in women, which was nearly identical to the previous study.15 Additionally, in both studies, the 95% LOA crossed zero, which indicates that BIA could either overestimate or underestimate LELM. However, in concordance with the previous study,15 we did not observe significant proportional bias, indicating that the magnitude of error was the same across all women, regardless of how much LELM they possessed. It is also worth noting that the BIA device used in the aforementioned study15 was from the same manufacturer as the ones used in the current study. Therefore, the agreement between our studies may have been enhanced by using the same prediction equation embedded in the devices.

We also noticed that the body composition profiles (LBM and LELM) of the women in our cohort were nearly identical to the athletes studied previously.15 Although our inclusion criteria only required participation in regular physical activity, the majority of our participants were National Collegiate Athletic Association Division I soccer players, which is similar to the cohort studied previously and explains the close similarity in body composition profiles. The consistency in results between our studies appears to lend credence to our findings regarding LELM assessments in women.

For men, we also observed underestimation bias (3.0 kg) of BIA, which is similar to a previous study of male collegiate football players, albeit smaller (5.4 kg).14 The magnitude of these systematic errors is much larger than an earlier study of 236 men between 18 and 30 years old in which they reported only a 0.2 kg difference between BIA and DXA; however, they did not report descriptive data for LELM so we cannot interpret the error relative to their body composition profiles. In the current study, we also noted wide LOA of approximately 4.2 kg, which is again smaller than the previous report (9 kg).14

When considering that the systematic error is larger in men than women and worsens when the amount of LELM increases, it appears that this may not be an issue of sex, but rather body composition profile. Because men naturally possess more lean mass than women, it appears that as LELM increases, so does the error, regardless of sex. This is supported by comparing our findings versus the cohort of collegiate football athletes, whereby those athletes possessed, on average, 31.0 ± 3.8 kg of LELM versus 23.1 ± 2.9 kg in the current study. The previous study also stratified their results by player position and found that the bias was smallest in wide receivers and largest in running backs14; these two positions recently have been shown to differ significantly in LELM.24 Collectively, it appears that there is a threshold at which the prediction equations for LELM cease to be valid. Because the algorithms embedded in BIA devices are proprietary to the manufacturer, we cannot make any suppositions as to where that threshold may be. However, because the algorithms apparently do not take sex into account, this lends more support to the error being related to the sheer magnitude of LELM.

LBM

Because BIA devices are constantly changing, advances in technology, validity, and reliability are specific to each device. Thus, we compared LBM between the BIA S10 (a relatively new and uninvestigated device) to DXA. We also included the data from the BIA 770 for comparative purposes between BIA devices to evaluate their interchangeability.

We observed excellent reliability and acceptable precision for both BIA devices when assessing LBM in men and women. Significant overestimation biases and wide LOA were noted only in women. These overestimation biases are nearly identical to previous studies.13,15 The small, non-significant overestimation bias we noted in our male cohort was similar in magnitude to one study,13 but stands in contrast to the data from football athletes that reported a significant overestimation error.14 Additionally, all three studies reported large ranges in the LOA, spanning approximately 6.4 kg (current study), 10.9 kg,13 and 17.7 kg,14 with all LOA ranges crossing zero.

Although the approximate 2 kg systematic differences we observed between BIA and DXA in women are reasonable, the broad LOA are of concern. For example, 8 kg of LBM (LOA range in women) represents 17.2% of the average LBM of the female cohort in the current study, which we would contend that most practitioners would consider clinically significant. We noted a proportional bias in women whereby the overestimation in total LBM was larger in those with more LBM, which is contrary to the results in female athletes15; however, because the investigators reported proportional bias as a trend from the Bland–Altman plots, it is unclear as to how our regression analysis of proportional bias compares. We did not find proportional bias in men, which indicates a “threshold” effect whereby the prediction equation is more valid for a particular body composition profile (ie, more LBM in men). Interestingly, a proportional bias was found in football players,14 indicating a potential inaccuracy when assessing a population whose body composition profile lies outside of the reference population. In the case of football players with large amounts of LBM, it is probable that BIA prediction equations do not extrapolate those values accurately. Future studies should attempt to validate the algorithms in a population of athletes with greater amounts of lean mass; however, this work may be in the hands of the manufacturers because the algorithms are proprietary.

Implications for Clinical Practice

We found significant discrepancies between BIA devices and DXA and wide LOA when measuring LELM, which were more pronounced in men than women. Based on our findings, it may be more acceptable to use BIA in individuals with lower amounts of LELM. However, because we do not know the threshold at which BIA becomes unacceptably inaccurate, we agree with previous authors that BIA could be appropriate for comparison within individuals (ie, tracking over time), but not comparison between individuals.14,15 This technology should not be used for diagnostic or prognostic criteria with regard to LELM. Although BIA measured LBM with a reasonable amount of systematic error, practitioners should be aware of the wide LOA that can result in an overestimation or underestimation and interpret accordingly.

References

  1. Moon JR. Body composition in athletes and sports nutrition: an examination of the bioimpedance analysis technique. Eur J Clin Nutr. 2013;67(S1)(suppl 1):S54–S59. doi:10.1038/ejcn.2012.165 [CrossRef]23299872
  2. Uhorchak JM, Scoville CR, Williams GN, Arciero RA, St Pierre P, Taylor DC. Risk factors associated with noncontact injury of the anterior cruciate ligament: a prospective four-year evaluation of 859 West Point cadets. Am J Sports Med. 2003;31(6):831–842. doi:10.1177/03635465030310061801 [CrossRef]14623646
  3. Evans KN, Kilcoyne KG, Dickens JF, et al. Predisposing risk factors for non-contact ACL injuries in military subjects. Knee Surg Sports Traumatol Arthrosc. 2012;20(8):1554–1559. doi:10.1007/s00167-011-1755-y [CrossRef]
  4. Bodkin SG, Werner BC, Slater LV, Hart JM. Post-traumatic osteoarthritis diagnosed within 5 years following ACL reconstruction [published online ahead of print March 18, 2019]. Knee Surg Sports Traumatol Arthrosc. doi:10.1007/s00167-019-05461-y [CrossRef]
  5. Brambilla L, Pulici L, Carimati G, et al. Prevalence of associated lesions in anterior cruciate ligament reconstruction: correlation with surgical timing and with patient age, sex, and body mass index. Am J Sports Med. 2015;43(12):2966–2973. doi:10.1177/0363546515608483 [CrossRef]26473010
  6. Going SB. Hydrodensitometry and air displacement plethysmography. In: Heymsfield SB, Lohman TG, Wang Z, Going SB, eds. Human Body Composition, 2nd ed. Champaign, IL: Human Kinetics; 2005:17–33.
  7. Levine JA, Abboud L, Barry M, Reed JE, Sheedy PF, Jensen MD. Measuring leg muscle and fat mass in humans: comparison of CT and dual-energy X-ray absorptiometry. J Appl Physiol (1985). 2000;88(2):452–456. doi:10.1152/jappl.2000.88.2.452 [CrossRef]
  8. Wang W, Wang Z, Faith MS, Kotler D, Shih R, Heymsfield SB. Regional skeletal muscle measurement: evaluation of new dual-energy X-ray absorptiometry model. J Appl Physiol (1985). 1999;87(3):1163–1171. doi:10.1152/jappl.1999.87.3.1163 [CrossRef]
  9. Norte GE, Knaus KR, Kuenze C, et al. MRI-based assessment of lower-extremity muscle volumes in patients before and after acl reconstruction. J Sport Rehabil. 2018;27(3):201–212. doi:10.1123/jsr.2016-0141 [CrossRef]
  10. Konrath JM, Vertullo CJ, Kennedy BA, Bush HS, Barrett RS, Lloyd DG. Morphologic characteristics and strength of the hamstring muscles remain altered at 2 years after use of a hamstring tendon graft in anterior cruciate ligament reconstruction. Am J Sports Med. 2016;44(10):2589–2598. doi:10.1177/0363546516651441 [CrossRef]27432052
  11. Ling CH, de Craen AJ, Slagboom PE, 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(5):610–615. doi:10.1016/j.clnu.2011.04.001 [CrossRef]21555168
  12. De Rui M, Veronese N, Bolzetta F, et al. Validation of bioelectrical impedance analysis for estimating limb lean mass in free-living Caucasian elderly people. Clin Nutr. 2017;36(2):577–584. doi:10.1016/j.clnu.2016.04.011 [CrossRef]
  13. Leahy S, O'Neill C, Sohun R, Jakeman P. A comparison of dual energy X-ray absorptiometry and bioelectrical impedance analysis to measure total and segmental body composition in healthy young adults. Eur J Appl Physiol. 2012;112(2):589–595. doi:10.1007/s00421-011-2010-4 [CrossRef]
  14. Raymond CJ, Dengel DR, Bosch TA. Total and segmental body composition examination in collegiate football players using multi-frequency bioelectrical impedance analysis and dual x-ray absorptiometry. J Strength Cond Res. 2018;32(3):772–782. doi:10.1519/JSC.0000000000002320 [CrossRef]
  15. Esco MR, Snarr RL, Leatherwood MD, et al. Comparison of total and segmental body composition using DXA and multifrequency bioimpedance in collegiate female athletes. J Strength Cond Res. 2015;29(4):918–925. doi:10.1519/JSC.0000000000000732 [CrossRef]
  16. McLester CN, Nickerson BS, Kliszczewicz BM, McLester JR. Reliability and agreement of various inbody body composition analyzers as compared to dual-energy x-ray absorptiometry in healthy men and women. J Clin Densitom. 2018;S1094-6950(18)30221-X. doi:10.1016/j.jocd.2018.10.008 [CrossRef]30472111
  17. Nickerson BS. Agreement between single-frequency bioimpedance analysis and dual energy x-ray absorptiometry varies based on sex and segmental mass. Nutr Res. 2018;54:33–39. doi:10.1016/j.nutres.2018.03.003 [CrossRef]29914665
  18. Lozano-Nieto A, Turner AA. Effects of orthostatic fluid shifts on bioelectrical impedance measurements. Biomed Instrum Technol. 2001;35(4):249–258.11494650
  19. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86(2):420–428. doi:10.1037/0033-2909.86.2.420 [CrossRef]18839484
  20. Denegar C, Ball D. Assessing reliability and precision of measurement: an introduction to intraclass correlation and standard error of measurement. J Sport Rehabil. 1993;2(2):35–42. doi:10.1123/jsr.2.1.35 [CrossRef]
  21. Koo TK, Li MY. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. J Chiropr Med. 2016;15(2):155–163. doi:10.1016/j.jcm.2016.02.012 [CrossRef]27330520
  22. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1(8476):307–310. doi:10.1016/S0140-6736(86)90837-8 [CrossRef]2868172
  23. Sergi G, De Rui M, Stubbs B, Veronese N, Manzato E. Measurement of lean body mass using bioelectrical impedance analysis: a consideration of the pros and cons. Aging Clin Exp Res. 2017;29(4):591–597. doi:10.1007/s40520-016-0622-6 [CrossRef]
  24. Bosch TA, Carbuhn AF, Stanforth PR, Oliver JM, Keller KA, Dengel DR. Body composition and bone mineral density of division 1 collegiate football players: a consortium of college athlete research study. J Strength Cond Res. 2019;33(5):1339–1346. doi:10.1519/JSC.0000000000001888 [CrossRef]

Mean ± SD for LELM and LBM from BIA Devices and DXA

Instrument Women (n = 31) Men (n = 26)
LELM (kg)
  DXA 16.7 ± 2.2 23.1 ± 2.9
  BIA 770 15.1 ± 2.4 19.9 ± 2.4
  BIA S10 15.3 ± 2.4 20.2 ± 2.4
LBM (kg)
  DXA 46.5 ± 5.3 66.4 ± 7.7
  BIA 770 48.9 ± 6.2 66.6 ± 8.0
  BIA S10 48.1 ± 6.1 65.7 ± 7.8

ICC and SEM (kg) for Agreement of LELM and LBM Results Between BIA Devices and DXA

Instrument Women (n = 31) Men (n = 26)
LELM (kg)
  BIA 770 0.82 (1.0) 0.69 (1.6)
  BIA S10 0.85 (0.9) 0.74 (1.2)
LBM (kg)
  BIA 770 0.93 (1.5) 0.99 (0.8)
  BIA S10 0.95 (1.3) 0.98 (1.0)

Bias and 95% LOA for LELM Compared to DXAa,b

Device Bias (kg) (P) 95% LOA (kg) (Range) Regression Coefficient (P)
Women (n = 31)
  BIA 770 1.6c (< .01) −0.9 to 4.0 (4.9) −0.31 (.35)
  BIA S10 1.3c (< .01) −1.1 to 3.7 (4.8) −0.37 (.27)
Men (n = 26)
  BIA 770 3.2c (< .01) 1.1 to 5.3 (4.2) 0.22d (< .01)
  BIA S10 2.9c (< .01) 0.7 to 5.0 (4.3) 0.18d (.03)

Bias and 95% LOA for LBM Compared to DXAa,b

Device Bias (kg) (P) 95% LOA (kg) (Range) Regression Coefficient (P)
Women (n = 31)
  BIA 770 −2.34c (< .01) −9.0 to −1.0 (7.9) −0.15d (.02)
  BIA S10 −1.54c (< .01) −8.2 to −0.2 (8.0) −0.14d (.04)
Men (n = 26)
  BIA 770 −0.16 (.63) −3.5 to 3.2 (6.7) −0.84 (.37)
  BIA S10 0.67 (.10) −3.2 to 4.5 (7.7) −0.27 (.73)
Authors

From the Department of Kinesiology, Center for Sport Performance, California State University–Fullerton, Fullerton, California (MMM); and the Department of Orthopaedics and Sports Medicine, University of South Florida, Tampa, Florida (AJT).

Supported in part by InBody USA, Cerritos, California.

The authors have no financial or proprietary interest in the materials presented herein.

Correspondence: Melissa M. Montgomery, PhD, ATC, 800 N. State College Blvd., Fullerton, CA 92831. E-mail: memontgomery@fullerton.edu

Received: July 03, 2019
Accepted: August 29, 2019
Posted Online: December 16, 2019

10.3928/19425864-20191024-01

Sign up to receive

Journal E-contents