Accurate measures of body composition (BC) are essential for performance and health. In addition to accuracy, BC measures should be practical and be minimally invasive to maximize their utility. The purpose of the present study was to compare the day-to-day variability and validity of four common laboratory-based body composition assessments to a criterion four-compartment model. Dual x-ray absorptiometry (DXA), air displacement plethysmography (BP), multi-frequency bioelectrical impedance (MF-BIA) and underwater weighing (UWW) were performed twice in a sample of 32 young men and women. Participants were assessed in a fasted, euhydrated state 2-7days apart. All methods were compared to a criterion four-compartment model using BP-derived body volume, DXA-derived bone mineral content and MF-BIA-derived total body water (4CBP ). Additional four-compartment models using UWW- and DXA-derived body volume were also examined (4CUWW ) and (4CDXA ). Validity results were conducted with paired t-tests and Bland-Altman analysis. Reliability was determined using intraclass correlations (ICC), coefficients of variation (CV) and standard error of the measurement (SEM). Validity analysis revealed that all methods overestimated per cent body fat and fat mass, and underestimated fat-free mass when compared with 4CBP , but only DXA and BP were significantly different (P<0·008). All measures were highly reliable across days (ICCs>0·9, CVs<12%). Results of the present study indicate that typical laboratory-based methods of body composition are valid and reliable. However, we caution that results should not be translated between methods and assessments should be performed with the same instrument when the goal is to monitor changes in body composition over time.
Read full abstract