Validity and agreement between dual-energy X-ray absorptiometry, anthropometry and bioelectrical impedance in the estimation of fat mass in young adults

Abstract

Assessment of fat mass has historically employed various methods like Dual-energy X-ray Absorptiometry (DXA), and bioelectrical impedance (BIA), and anthropometry with its set of formulas. However, doubts persist regarding their validity and interchangeability to evaluate fat mass. This research aimed to determine the validity of anthropometry, and BIA in estimating fat mass Vs DXA, considering the influence of sex and hydration status.A descriptive, cross-sectional study included 265 young adults (161 males and 104 females), assessed through DXA, BIA in a standing position, and anthropometry. A fat mass estimation formula with DXA, a fat mass estimation formula with BIA and 10 fat mass estimation formulas with anthropometry were calculated.Significant differences were found across DXA, BIA and anthropometry in both kilograms and percentages for the overall sample (p<0.001), and when the covariable sex was included (p<0.001), with no significant effect of hydration status (p=0.332-0.527). Bonferroni-adjusted analyses revealed significant differences from DXA with anthropometry and BIA in most cases for the overall sample (p<0.001), as well as when stratified by sex (p<0.001–0.016). Lin’s coefficient indicated poor agreement between most of the formulas and methods both in percentage and kilograms of fat mass (CCC=0.135–0.892). In the Bland-Altman analysis, using the DXA fat mass values as a reference, lack of agreement was found in the general sample (p<0.001-0.007), except for Carter’s formula in kilograms (p=0.136) and percentage (p=0.929) and Forsyth for percentage (p=0.365). When separating the sample by sex, lack of agreement was found in males for all methods when compared with both percentage and kilograms calculated by DXA (p<0.001). In the female sample, all methods and formulas showed lack of agreement (p<0.001–0.020), except for Evans’s in percentage (p=0.058).The formulas for fat mass assessment with anthropometry and BIA may not be valid with respect to the values reported with DXA, with the exception of Carter’s anthropometry formula for general sample and Evans’s anthropometry formula for female sample. BIA could also be an alternative if what is needed is to assess fat mass in women as a group.