Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Nov;63 Suppl 2(Suppl 2):S35-S47.
doi: 10.1002/jcph.2306.

Estimation of Absolute and Relative Body Fat Content Using Noninvasive Surrogates: Can DXA Be Bypassed?

David J Greenblatt  1   2 , Christopher D Bruno  1   3 , Jerold S Harmatz  1 , Bess Dawson-Hughes  4 , Qingchen Zhang  1 , Chunhui Li  1 , Christina R Chow  3
Affiliations

Estimation of Absolute and Relative Body Fat Content Using Noninvasive Surrogates: Can DXA Be Bypassed?

David J Greenblatt et al. J Clin Pharmacol. 2023 Nov.

Abstract

Dual-energy x-ray absorptiometry (DXA) scanning is used for objective determination of body composition, but instrumentation is expensive and not generally available in customary clinical practice. Anthropometric surrogates are often substituted as anticipated correlates of absolute and relative body fat content in the clinical management of obesity and its associated medical risks. DXA and anthropometric data from a cohort of 9230 randomly selected American subjects, available through the ongoing National Health and Nutrition Examination Survey, was used to evaluate combinations of surrogates (age, height, total weight, waist circumference) as predictors of DXA-determined absolute and relative body fat content. Multiple regression analysis yielded linear combinations of the 4 surrogates that were closely predictive of DXA-determined absolute fat content (R2 = 0.93 and 0.96 for male and female subjects). Accuracy of the new algorithm was improved over customary surrogate-based predictors such as body mass index. However prediction of relative body fat was less robust (R2 less than 0.75), probably due to the nonlinear relation between degree of obesity (based on body mass index) and relative body fat. The paradigm was validated using an independent cohort from the National Health and Nutrition Examination Survey, as well as two independent external subject groups. The described regression-based algorithm is likely to be a sufficiently accurate predictor of absolute body fat (but not relative body fat) to substitute for DXA scanning in many clinical situations. Further work is needed to assess algorithm validity for subgroups of individuals with "atypical" body construction.

Keywords: National Health and Nutrition Examination Survey (NHANES); anthropometric surrogates; body fat content; dual-energy x-ray absorptiometry (DXA); epidemiology; obesity.

PubMed Disclaimer

Conflict of interest statement

Conflicts of Interest

CDB and CRC are employees of Emerald Lake Safety, LLC. Other authors do not have conflicts of interest.

Figures

Figure 1.
Figure 1.
Plots of mean (with standard deviation) values of weight, height, waist circumference, and body mass index in relation to age among male and female subjects in the NHANES 2003–2006 cohort. Points represent the mean value for all subjects within the indicated 10-year age interval, plotted against the midpoint of the age interval.
Figure 2.
Figure 2.
Plots of mean (with standard deviation) percent ideal body weight and percent body fat (by DXA) in relation to age among male and female subjects in the NHANES 2003–2006 cohort. Points represent the mean value for all subjects within the indicated 10-year age interval, plotted against the midpoint of the age interval.
Figure 3.
Figure 3.
Relation of measured total fat by DXA (Y-axis) to body mass index (X-axis) among male and female subjects in the NHANES 2003–2006 cohort. Solid lines are the functions determined by linear regression analysis.
Figure 4.
Figure 4.
Relation of fraction of total weight comprised of fat (by DXA) (Y-axis) to waist circumference (top 2 panels) and body mass index (bottom 2 panels) (X-axes) among male and female subjects in the NHANES 2003–2006 cohort. Solid lines are the functions determined by nonlinear regression analysis using a function of the form shown in Equation 3. The maximum (max) value of fraction total fat (0.48 for males, 0.55 for females) is based on the fitted function (shown as Ymax in Equation 3).
Figure 5.
Figure 5.
Relation of measured total fat by DXA (Y-axis) to predicted values of total fat based on the multiple regression algorithm shown in Table 4 (X-axis) among male and female subjects in the NHANES 2003–2006 cohort. Dashed lines are the lines of identity (Y = X), which could not be visually distinguished from the functions determined by linear regression analysis.
Figure 6.
Figure 6.
Frequency distributions of the difference (in percent) between algorithm-predicted total fat and measured total fat (by DXA) (Equation 1) among male and female subjects in the NHANES 2003–2006 cohort. Lines represent the functions determined by nonlinear regression analysis based on a normal distribution. X¯ values above the arrows are the overall arithmetic means.
Figure 7.
Figure 7.
Relation of measured fraction of total fat measured by DXA (Y-axis) to predicted values of fractional total fat based on the multiple regression algorithm shown in Table 4 (X-axis) among male and female subjects in the NHANES 2003–2006 cohort. Dashed lines are the lines of identity (Y = X), which could not be visually distinguished from the functions determined by linear regression analysis.
Figure 8.
Figure 8.
Relation of measured total fat by DXA among male and female subjects in the Emerald Lake Safety LLC study cohort (Table 2) (Y-axis) to predicted values of total fat based on anthropometric values in that cohort, using the multiple regression algorithm from the 2003–2006 NHANES model development cohort (X-axis), as shown in Table 4. Solid line is the function determined by linear regression analysis. Dashed line is the line of identity (Y = X).
Figure 9.
Figure 9.
Relation of measured total fat by DXA among male and female subjects in the STOP-IT study cohort (Table 2) (Y-axis) to predicted values of total fat based on anthropometric values in that cohort, using the multiple regression algorithm from the 2003–2006 NHANES model development cohort (X-axis), as shown in Table 4. Dashed line is the line of identity (Y = X). Regression lines are not shown.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Ward ZJ, Bleich SN, Cradock AL, et al. Projected U.S. State-Level Prevalence of Adult Obesity and Severe Obesity. N Engl J Med. 2019; 381(25): 2440–2450. - PubMed
    1. Ellison-Barnes A, Johnson S, Gudzune K. Trends in Obesity Prevalence Among Adults Aged 18 Through 25 Years, 1976–2018. JAMA. 2021; 326(20): 2073–2074. - PMC - PubMed
    1. Williamson K, Nimegeer A, Lean M. Rising prevalence of BMI ≥ 40 kg/m2 : A high-demand epidemic needing better documentation. Obes Rev. 2020;21(4):e12986. - PMC - PubMed
    1. Hruby A, Hu FB. The Epidemiology of Obesity: A Big Picture. Pharmacoeconomics. 2015; 33(7): 673–689. - PMC - PubMed
    1. Blüher M Obesity: global epidemiology and pathogenesis. Nat Rev Endocrinol. 2019; 15(5): 288–298. - PubMed

Publication types

LinkOut - more resources