Comparison of anthropometry to DXA: a new prediction equation for men

Eur J Clin Nutr. 2004 Nov;58(11):1525-31. doi: 10.1038/sj.ejcn.1602003.


Objective: This study compared three professionally recommended anthropometric body composition prediction equations for men to dual energy X-ray absorptiometry (DXA), and then developed an updated equation, DXA Criterion (DC) from DXA.

Design: Cross-sectional.

Setting: Exercise Physiology Lab. University of Missouri-Columbia, USA.

Subjects: A total of 160 men aged 18-62 y old.

Interventions: Percent body fat (%BF) by anthropometry was compared to DXA on the same day.

Results: Although %BF was significantly correlated (r=0.923-0.942) (P<0.01) with DXA for all three equations, each equation underestimated %BF (range=3.1-3.3%) (P<0.01) compared to DXA. The following DC equation for men was created: %BF=0.465+0.180(Sigma7SF)-0.0002406(Sigma7SF)(2)+0.06619(age); (Sigma7SF=sum of chest, midaxillary, triceps, subscapular, abdomen, suprailiac, thigh; age=years). The predicted residual sum of squares (PRESS) R(2) was high (0.90) and the PRESS standard error of estimates was excellent (2.2% at the mean) for the DC equation when applied to our sample of 160 men.

Conclusions: The currently recommended anthropometric equations for men underestimate %BF compared to DXA. The DC equation yields a more accurate estimation of %BF in men aged 18-62 y old. The results from this study support the need for the current %BF standards and norms for men to be adjusted upward.

Publication types

  • Comparative Study

MeSH terms

  • Absorptiometry, Photon / methods*
  • Adipose Tissue / anatomy & histology*
  • Adolescent
  • Adult
  • Anthropometry / methods*
  • Body Composition*
  • Cross-Sectional Studies
  • Humans
  • Male
  • Mathematics
  • Middle Aged
  • Predictive Value of Tests
  • Reproducibility of Results
  • Sensitivity and Specificity