Travis M. Combest, MS, MPH*; Robin S. Howard, MA ; LTC Anne M. Andrews, SP USA (Ret.)

Transcription

1 MILITARY MEDICINE, 182, 7/8:e1908, 2017 Comparison of Circumference Body Composition Measurements and Eight-Point Bioelectrical Impedance Analysis to Dual Energy X-Ray Absorptiometry to Measure Body Fat Percentage Travis M. Combest, MS, MPH*; Robin S. Howard, MA ; LTC Anne M. Andrews, SP USA (Ret.) ABSTRACT Background: To compare the circumference measurement (CM) body composition method and 8-point segmental bioelectrical impedance analysis (DSM-BIA) to dual energy X-ray absorptiometry (DEXA) in military members. Objective: The objective was to compare three body composition methods. Our hypothesis was the CM is as accurate as DSM-BIA and DEXA in assessing body fat percentage (%BF). Design: Cross sectional, observational study. Participants/Setting: Healthy active duty military males and nonpregnant females. Seventy-six participants (mean age 35.0 ± 9.7 years, mean body mass index 28.9 ± 4.7 kg/m 2 ), outpatient clinic setting was used. Statistical Analyses: Agreement between DEXA and the other two methods was examined using an intraclass correlation coefficient (ICC) using the two-way random method with absolute agreement. Repeated measures analysis of covariance was used to examine the effect of gender and waist circumference on differences in %BF. Results: The agreement of DSM-BIA with DEXA for females was ICC = 0.93 (95% confidence interval [CI]: ) and for males, ICC = 0.89 (95% CI: ). For the agreement of CM with DEXA, the ICC for females was 0.83 (95% CI: ) and for males the ICC = 0.72 (95% CI: ). For females with smaller waists (<81.3 cm), the DSM-BIA underestimated the DEXA measurement by a mean of 1.6% (95% CI: %). For males with larger waists ( 95.3 cm), the DSM- BIA overestimated the DEXA measurement by a mean of 2.6% (95% CI: %). For females with larger waists ( 81.3 cm), the CM overestimated the DEXA %BF by an average of 2.4% (95% CI: %). Conclusion: There was good agreement between the three methods assessed in this study. Both waist circumference and gender had an effect on the accuracy of the DSM-BIA and CM measurements. INTRODUCTION The increased prevalence of obesity and cardiovascular disease, where obesity is a significant risk factor, has made the measurement of body fat increasingly applicable to the U.S. military and the overall health of military service members (SMs). 1 Furthermore, excessive body fat can decrease anaerobic and aerobic activity capabilities and increase injury risks. 2 During the time period of , the medical surveillance monthly report found that there were a total of 382,448 active component SMs were found to be overweight during a medical encounter. 1 Because of the large population and the need for standardization, a preferred method of measuring body fatness should be simple, portable, and reliable. Circumference measurement (CM) is one such method that uses a tape measure to measure specific anatomical sites. CMs are routinely obtained in accordance with Department of Defense (DoD) regulations and for *Department of Nutrition Services, Walter Reed National Military Medical Center, Bethesda, MD Department of Research Programs, Walter Reed National Military Medical Center, Bethesda, MD Department of Rehabilitation, Walter Reed National Military Medical Center, Bethesda, MD Human Subjects Protection Office, National Standards and Technology, Gaithersburg, MD The views expressed in this presentation are those of the author(s) and do not reflect the official policy of the Department of Army, Navy, Defense, or any other agency of the U.S. Government, nor should any official endorsement be inferred by the Department of Defense, or the U.S. Government. doi: /MILMED-D all SM who exceed the weight-for-height screening requirements. 3 7 The measurements are entered into a calculation, along with height, to calculate body fat percentage (%BF). The accuracy and reliability of %BF results using CM are highly dependent on operator performance, which may introduce error between raters. 8 The Standard Error of Estimate for males was 3.52% BF and for females was 3.61% BF. 9 Dual-energy X-ray absorptiometry (DEXA) is now considered the gold standard for measuring body composition. 10 In recent years, utilization of DEXA for assessing body composition has become more prevalent than underwater weighing because of its ease of use, minimal patient discomfort, beneficial segmental analysis, ability to account for bone mineral density, and results are consistent with underwater weighing DEXA is used to assess U.S. citizens nutritional and health status in the National Health and Nutrition Examination Survey. 10 DEXA measures bone mineral content, lean tissue mass, and fat-mass, and the precision error for each is 1.8%, 1.5%, and 1.5% respectively. 13 Limits to DEXA include immobility, increased expense, exposition to low-grade radiation, and the size of the scanning area. 12,13 Bioelectrical Impedance Analysis (BIA) is another method utilized to estimate %BF. BIA is easy to administer with little patient discomfort and it is less expensive than DEXA, portable and requires minimal maintenance. BIA, as with DEXA, takes weight and height into consideration. Three components differentiate between models: body segmentation, frequency (khz), and empirical equations. Body e1908

2 segmentation refers to how BIA is used to assess the body, as a one whole cylindrical rod or in five cylindrical segments that are broken down by each limb and one trunk Usual currents are 5, 50, 250, and 500 khz and multifrequencies use any combination of currents. 16,17 Multifrequency models use low frequencies to detect extracellular water and high frequencies to measure intracellular water. 16,17 Factors such as hydration and obesity can cause the BIA measurement to be less accurate when compared to DEXA. 16,17 Compared to DEXA, the Standard Error of Estimate for males is 1.76 and for females is Recent adoptions of the DoD %BF standards by multiple military services warrants an updated comparison with other body composition methods. The purpose of this study was to compare three different body composition methods for assessing %BF in active duty SM. We hypothesized that (1) DSM-BIA and CMs are as accurate as DEXA in assessing %BF in active duty SMs and (2) hydration has no effect on the difference of measured %BF outcomes of the DSM-BIA. SUBJECTS AND METHODS Subjects Male and female SM were recruited using flyers posted around the installation and by electronic mail to staff members at Walter Reed Army Medical Center. Subjects were excluded if they were pregnant, taking electrolyte altering medications, or had metal-containing materials in their body. The Institutional Review Board of Walter Reed Army Medical Center approved the study. All subjects provided written informed consent after receiving verbal and written information about the study. Data collection for each subject was performed in one visit. Anthropometric Measurements Height was recorded to the nearest 0.5 in and converted to centimeter using a Stadiometer (Seca, Hamburg, Germany). Weight without shoes was measured to the nearest 0.1 kg using a digital scale (Seca) or the 8-point segmental bioelectrical analysis referred to as Direct Segmental Multifrequency Bioelectrical Impedance Analysis (DSM-BIA) digital built-in scale. Both scales were calibrated with no significant difference between instruments. Body mass index was calculated: weight (kg)/height (m 2 ). CMs were measured in inches in accordance with AR 600-9, which reflects DoD Instruction for males (neck and waist) and females (neck, waist, and hips). 4 A nonstretching tape was used to measure each body landmark three times with sufficient tension to measure the site without indenting skin surface. All tape placements were parallel to the floor and read at eye level. The average of three measurements from each anatomical site was rounded up or down to the nearest 0.5 in and converted to centimeter. The neck site was measured right below the larynx, with the subject standing up and facing forward with nonhunched shoulders, and the measurement was rounded up. The waist circumference (WC) was gender specific, and the measurement was rounded down. Males were taped at the navel, and females were taped at the natural waistline or the smallest WC usually found halfway between the naval the sternum. The hip measurement taken at the largest protrusion of the buttocks was performed solely on females. Body fat percentage was assessed using a whole-body scan with a DEXA (Hologic Inc QDR, Bedford, Massachusetts). DEXA uses fan-beam X-ray scanning to accomplish a threecompartment model that differentiates between fat mass, fatfree mass, and bone mineral content. Two different X-ray radiation energies penetrate the body, and the corresponding attenuation levels are used to compute %BF. 13,19 Daily calibration for DEXA was performed in accordance with manufacturer specifications. The same DEXA was used to measure all subjects, and the primary investigator reviewed all DEXA results for accuracy. Subjects wore loose-fitting clothing and no shoes. All jewelry, piercings, and metal hair items were removed before the scan. All subjects assumed a supine position during the scanning process and were instructed to remain still and refrain from talking, while the scan was in progress. Total scanning time was 10 minutes or less. Body fat percentage was also assessed by using the Inbody point DSM-BIA (Biospace Inc., Los Angeles, California). The DSM-BIA uses the three-compartment model and segmental analysis that differentiates between fat mass, fat-free mass, and water mass. The Inbody point DSM-BIA uses alternate currents of 5, 50, and 500 khz to penetrate all body tissue types. 14 The same DSM-BIA instrument was used to measure all subjects. Subjects stood barefoot on marked footprints on the DSM-BIA machine. Two electrodes were in contact by each foot and hand through the base and the handgrips on the instruments. During the assessment, subjects stood completely still with their arms away from their torsos and slightly bowed. Scanning took less than 1 minute to complete. Urine specific gravity, using a calibrated refractometer, was used to assess hydration status. Urine samples were collected from each subject during the assessment process. A small sample of urine was applied to the refractometer using a cotton swab. Urine specific gravity was reported to the 0.01 SG units. STATISTICAL METHODS Body fat percentage was measured by all three methods (DEXA, DSM-BIA, and CM), with DEXA as the reference standard. Data are presented as the mean ± SD or as means with 95% confidence intervals (95% CIs). Agreement between DEXA and the other two methods was examined using an intraclass correlation coefficient (ICC) using the two-way random method with absolute agreement. Repeated measures analysis of covariance was used to examine differences in %BF between the three methods controlling for e1909

3 gender and WC. To present %BF according to WC for each gender, a median split was used: The median WC for each sex (95.3 cm for males and 81.3 cm for females) was used as a cut point to create two groups within each gender. A paired t test was used to compare %BF within each gender and WC group, and p values were adjusted using a Bonferroni correction. The association of WC with the error between DEXA and DSM-BIA or CM is described using a Pearson correlation coefficient. All analyses were performed using SPSS Version 21 (SPSS Inc., Chicago, Illinois). A p value <0.05 was considered statistically significant. As per the protocol, for estimation of the minimum sample size, the probability of a Type 1 error was controlled at alpha = (the experiment-wise alpha error of 0.05 was reduced using a Bonferroni correction for three possible pairwise comparisons between DEXA, DSM-BIA, and CM) and a SD of 4.1% was used on the basis of differences in the literature between DEXA and DSM-BIA. 20 A sample of 50 subjects provided 80% power to detect a 2% difference between the measures. RESULTS Subject characteristics are summarized in Table I. Seventysix subjects (31 males and 45 females) participated in the study and completed all three body fat measurement methods. Mean hydration status for all subjects indicated that all subjects were euhydrated. In terms of overall agreement with DEXA, the ICC for DSM-BIA for females was 0.93 (95% CI: ) and for males, ICC = 0.89 (95% CI: ). For the agreement of CM with DEXA, the ICC for females was 0.83 (95% CI: ) and for males the ICC = 0.72 (95% CI: ). Differences between DEXA and DSM-BIA in females ranged from 7.7% to +6.9% (mean 0.04%, 95% CI: %). In males, this difference ranged from 6.8% to +4.7% (mean 0.8%, 95% CI: %) (Fig. 1). For both females and males, these differences were significantly associated with WC (r = 0.53, p < 0.001). Differences between DEXA and CM in females ranged from 12.3% to +7.4% (mean 1.4%, 95% CI: 2.6% to 0.1%). In males, TABLE I. Subject Characteristics Male (n = 31) Female (n = 45) p Values Age (Years) 32.7 ± ± Weight (kg) 93.3 ± ± 14.5 <0.001 BMI (m/kg 2 ) 29.4 ± ± Neck (cm) 40.8 ± ± 2.2 <0.001 Hips (cm) a N/A ± 10.1 Waist (cm) 95.2 ± ± 9.6 <0.001 SG b 1.02 ± ± FIGURE 1. Differences between DSM-BIA and DEXA according to gender and WC. this difference ranged from 7.7% to +10.3% (mean 1.3%, 95% CI: %) (Fig. 2). For females, the differences between DEXA and CM measurements were negatively associated with WC (r = 0.36, p = 0.015) but there was no significant association for males (r = 0.10, p = 0.60). All three measurement methods of %BF were compared in a multivariate model accounting for sex and WC. There was a statistically significant difference in %BF between methods ( p = 0.001), gender ( p < 0.001), and WC ( p < 0.001). There was also a significant interaction between the methods and gender ( p = 0.001), indicating that the measurement methods did not measure %BF similarly between males and females. Measurement of body fat percent wasthenexaminedineachgenderseparately.forbothmales and females, there was a significant interaction between the methods and WC ( p = for males, p = for females) indicating that within each gender, measurement methods produced different results according to WC. Body fat measurements by gender and WC are presented in Table II. For females with smaller waists (<81.3 cm), the DSM-BIA underestimated the DEXA measurement by a BMI, body mass index. Data are mean ± SD. a Hip measurements were not measured in males and were not compared with females. b Hydration status all subjects n = 75, male n = 30. FIGURE 2. and WC. Differences between CM and DEXA according to gender e1910

4 TABLE II. Body Fat Measurements According to Gender and WC Body Fat (%) p Values a Group n DEXA BIA CM DEXA vs. BIA DEXA vs. CM Males ± ± ± 5.1 Waist < 95.3 cm ± ± ± Waist 95.3 cm ± ± ± Females ± ± ± 7.9 Waist < 81.3 cm ± ± ± Waist 81.3 cm ± ± ± Data are mean ± SD. a Methods compared using a paired t test with p values adjusted using a Bonferroni correction. mean of 1.6% (95% CI: %). For males with larger waists ( 95.3 cm), the DSM-BIA overestimated the DEXA measurement by a mean of 2.6% (95% CI: %). For females with larger waists ( 81.3 cm), the CM overestimated the DEXA %BF by an average of 2.4% (95% CI: %). DISCUSSION The CM measures were lower than the DEXA measures in males, but it did not reach statistical significance. The CM method, however, overestimates %BF in females with a WC >81.3cm when compared with DEXA. Bathalon et al 21 explored differences between two CM equations (DoD method using the sites of neck, waist, and hip vs. previous U.S. Army method using the neck, wrist, and thigh) and its effect on measurements for female SMs. The updated DoD method did not significantly change the number of females who were categorized as noncompliant. For females, high WC was an indicator of decreased DEXA-CM agreement. For males, high WC was an indicator of decreased DEXA-DSM-BIA agreement. The use of CM, specifically WC, may be used as a health indicator to promote early medical involvement to address type II diabetes or/and heart disease. Bathalon et al 22 completed a record review of Soldiers enrolled in the Army Weight Control Program and results indicated that up to 30% of their sample had one or more elevated lab values associated with disease such as: cholesterol, triglycerides, and/or glucose levels The Army Weight Control Program was able to assess this at-risk population because of the %BF standards that were set in place by the military and subsequent screening protocols. In previous research, BIA had great variance and varied agreement with DEXA. Depending on the type of BIA method used (multifrequency, whole body, upper body, DSM), its agreement with DEXA can vary, ranging from r =0.78to0.85formenandwomen,respectively. 23 The usage of BIA for testing body composition is very quick, inexpensive, and requires little training for test administrators. We found no difference in %BF results when compared to DEXA, indicating that this technique may be ideal for fitness centers, sports medicine and weight reduction programs, and other high volume venues. 24 The BIA method has been shown to over- and underestimate %BF as a result of hydration status. We did not find that hydration status was a significant confounding factor in our study analysis. Duz et al 25 concluded BIA %BF predictions differed from DEXA with exercise, fluid intake, dehydration, and smoking using a handheld BIA model (Single-frequency, two electrode, equation based). The DEXA-BIA mean difference in Duz et al 25 study was 7.0 ± 0.6 %BF, which is higher than our study results 0.9 ± 2.2 % BF (males: 0.8 ± 1.5 %BF and females: 0.1 ± 1.7 %BF) using a DSM-BIA. Our study only accounted for hydration status, and assumed subjects did not exercise 4 hours before data collection per protocol instructions. Duz et al 25 also stated that in general, BIA underestimated %BF when compared to DEXA but overestimated %BF in subjects who were hypohydrated. Our subjects were euhydrated, which may indicate that if subjects are hydrated, they may have similar nonsignificant results between DSM-BIA and DEXA. Our findings suggest that DSM-BIA tends to overestimate %BF as WC increases. Large WC results in DSM-BIA overestimated %BF when compared to DEXA. In males who had >95.3 cm waist the DSM-BIA was 1.1% less than DEXA but was not found to be significant. In females with a waist <81.3 cm, there was a significant difference resulting in a 1.5% less %BF than DEXA. Shafer et al 26 found that the significant predictor for error between DEXA and segmental, multiple-frequency BIA %BF was WC. Other studies have had similar findings in that WC has been shown to be a significant factor in disagreement between DEXA and DSM-BIA. As technology evolves and BIA analyzers are capable of higher frequencies, we suspect the error will improve. This study was strengthened by each subject serving as their own control and all measures were completed in one session. The study would have been strengthened by environmental controls such as time of day and verification of exercise and fluid intake before testing. However, these were not feasible in the research setting. CONCLUSION This study found no significant difference between the three methods for all subjects, but that there were important e1911

5 differences noted when gender and WC were included in the analyses. Reliable methods to estimate %BF that are accessible to practitioners are essential given the importance evaluating body fatness for chronic disease risk and for military administrative purposes. This study demonstrated that CM and DSM-BIA methods are valid for these purposes. Future studies should continue to explore the effects of other variables such as race and age on the accuracy of body composition measurements. In addition, more research should focus on how to mitigate or control for the effect of WC on these methods. ACKNOWLEDGMENTS We would like to respectfully acknowledge all subjects for their participation in the research. The study was supported by Walter Reed Army Medical Center Department of Research Programs, Washington, DC. REFERENCES 1. Brundage J (editor). Diagnoses of overweight/obesity, active component, U.S. Armed Forces, MSMR 2011; 18(1): Crawford KF, Abt J, Sell T, et al: Less body fat improves physical and physiological performance in army soldiers. Mil Med 2011; 176(1): Headquarters, Department of the Army. Army regulation 600-9: The Army weight control program. Washington, DC. Revised June 28, Effective July 28, Available at Search/ePubsSearch/ePubsSearchDownloadPage.aspx?docID=0902c ; accessed August 1, Department of Defense. DoD Instruction number : Physical fitness and body fat programs procedures. November 5, Available at accessed on August 1, Chief of Naval Operations, Department of the Navy. OPNAV Instruction H: Physical readiness program. Washington, DC. August 15, Available at h/6110.1H.pdf; accessed August 1, Commandant of the Marine Corps, Department of the Navy. MCO : Marine Corps body composition and military appearance program. Washington, DC. August 8, Available at pdf; accessed August 1, United States Coast Guard, U.S. Department of Homeland Security. Commandant Instruction M1020.8G: Coast Guard weight and body fat standards program manual. Washington, DC. June 30, Available at accessed August 1, Nagy E, Vriendt TD, Ortega FB, et al: Harmonization process and reliability assessment of anthropometric measurements in a multicenter study in adolescents. Int J Obes 2008; 32: s Hogdon J, Friedl KE: Development of the DoD Body Composition Estimation Equations. Technical Document No. 99-2B. San Diego, CA. Naval Health Research Center, Available at: UKEwjPiuuZq4HTAhXCMSYKHem3DvYQFggvMAM&url=http%3A %2F%2Fwww.dtic.mil%2Fcgi-bin%2FGetTRDoc%3FAD%3DADA &usg=AFQjCNHfIKuia6ATD2qSQLrO5fXeD1DJYA&bvm=bv ,d.eWE; accessed November 17, Li C, Ford ES, Zhao G, Balluz LS, Giles WH: Estimates of body composition with dual-energy X-ray absorptiometry in adults. Am J of Clin Nutr 2009; 90(6): Shaw KA, Srikanth VK, Fryer JL, Blizzard L, Dwyer T, Venn AJ: Dual energy X-ray absorptiometry body composition and aging in a population-based older cohort. Int J Obes 2007; 31: Williams JE, Wells JC, Wilson CM, Haroun D, Lucas A, Fewtrell MS: Evaluation of Lunar Prodigy dual-energy X-ray absorptiometry for assessing body composition in healthy persons and patients by comparison with the criterion 4-component model. Am J Clin Nutr 2006; 83(5): Lohman M, Tallroth K, Kettunen JA, Marttinen MT: Reproducibility of dual-energy x-ray absorptiometry total and regional body composition measurements using different scanning positions and definitions of regions. Metabolism 2009; 58(11): Biospace America, Inc. Inbody Technology. Available at accessed on August 1, Ishiguro N, Kanehisa H, Miyatani M, Masuo Y, Fukunaga T: Applicability of segmental bioelectrical impedance analysis for predicting trunk skeletal muscle volume. J Appl Physiol 2006; 100(2): Biospace America, Inc.: Inbody 570 Pinnacle Performance. Available at accessed August 1, Medici G, Mussi C, Fantuzzi AL, Malavolti M, Albertazzi A, 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(8): Demura S, Sato S, Kitabayashi T: Percentage of total body fat as estimated by three automatic bioelectrical impedance analyzers. J Physiol Anthropol Appl Human Sci 2004; 23(3): Yamada Y, Masuo Y, Yokoyama K, et al: Proximal electrode placement improves the estimation of body composition in obese and lean elderly during segmental bioelectrical impedance analysis. Eur J Appl Physiol 2009; 107(2): Pietrobelli A, Tatò L: Body composition measurements: from the past to the future. Acta Paediatrica 2005; 94: Sun G, French CR, Martin GR, et al: Comparison of multifrequency bioelectrical impedance analysis with dual-energy X-ray absorptiometry for assessment of percentage body fat in a large, healthy population. Am J Clin Nutr 2005; 81(1): Bathalon GP, Sharp MA, Williamson DA, Young AJ, Friedl KE: The effect of proposed improvements to the Army weight control program on female Soldiers. Mil Med 2006; 171: Bathalon GY, Andrew JF, Worley MA, Kinsey RF, Friedl KE: Assessment of excess weight and fat in Army weight control program participants. Am J Sports Med [Poster Presentation] 2006; 38(5): S Frisard MI, Greenway FL, Delany JP: Comparison of methods to assess body composition changes during a period of weight loss. Obes Res 2005; 13(5): Duz S, Kocak M, Korkusuz F: Evaluation of body composition using three different methods compared to dual-energy X-ray absorptiometry. Eur J Sport Sci 2009; 9(3): Shafer KJ, Siders WA, Johnson LK, Lukaski HC: Validity of segmental multiple-frequency bioelectrical impedance analysis to estimate body composition of adults across a range of body mass indexes. Nutrition 2009; 25(1): e1912