To identify which equation, Siri or Brozek, based on the two compartment model, provides a more accurate conversion of body density (BD) in percent body fat (%BF) in a group of older adults.

Cross-sectional study.

Research Centre in Physical Activity, Health and Leisure, Faculty of Sport, University of Porto.

60 older adults, aged 60–92 years.

Skinfold thickness was used to estimate BD through Visser et al. prediction equation. The conversion of BD to %BF was done with Siri (%BF-Siri) and Brozek (%BF-Brozek) formulas and these determined values were both compared to Dual-Energy X-ray Absorptiometry (%BF-DXA) evaluations.

A strong correlation between the %BF-DXA value and %BF-Siri (r=0.91, p<0.001) and %BF-Brozek (r=0.91, p<0.001) was found, although %BF-Siri and %BF-Brozek overestimated %BF-DXA (p<0.001). The comparison of the %BF-Siri and %BF-Brozek mean values also revealed significant differences (p<0.001). The %BF-Brozek reflects a better agreement than the %BF-Siri with %BF-DXA with respectively a mean difference of −4.0%BF (limits of agreement = −10.9 to 2.9%) and −5.7%BF (−12.6 to 1.2). The Bland and Altman plots confirmed that %BF-Brozek reflects a better agreement with %BF-DXA.

The results of the present study show that the use of Brozek equation may correspond to a more accurate alternative than Siri equation for the conversion of BD in %BF in older adults.