Evaluation of dual energy X-ray absorptiometry as a method of measurement of body fat

Two densitometers, the Hologic QDR 1000W densitometer (QDR) and the Lunar DPX densitometer (DPX), were compared with four traditional methods for the measurement of body fat: underwater weighing (UWW), skinfold thickness measurements (SKM), bioelectrical impedance analysis (BIA) and deuterium oxide dilution (D2O). Precision was assessed by conducting 10 repeated observations on one subject using QDR: the coefficient of variability (CV) was 1.8% for percentage fat, 0.6% for lean mass and 2.1% for fat mass. In 10 repeated observations on three subjects, QDR CV was greater than UWW (CV percentage fat QDR 1.3% (range 0.9-1.6%) compared with 4.8% (range 3.8-6.6%) for percentage fat UWW). Based on observations of 12 subjects, correlations of QDR and DPX with UWW for percentage fat were high: Pearson's r = 0.916, P < 0.0001 for QDR and r = 0.913, P < 0.00001 for DPX. A limits of agreement estimate showed a between-method difference of +1.3% (range -4% to +7%) for QDR compared with UWW. The DPX showed a between-method difference of +4.8% (range +2% to +9%) compared with UWW. The correlation of the two DEXAs was high (r = 0.986, P < 0.0001). Correlations between both DEXA instruments and other methods were high (for QDR: r = 0.824, P < 0.001 for SKM; r = 0.972, P < 0.0001 for BIA; r = 0.787, P < 0.002 for D2O; for DPX: r = 0.923, P < 0.00001 for SKM; r = 0.910, P < 0.00001 for BIA; r = 0.812, P < 0.001 for D2O). It was concluded that QDR and DPX measured percentage fat with greater precision than UWW as reflected by the CV and correlated with other methods. Predicted percentage fat measurements using QDR would be 3% lower than DPX, but both DEXAs predict percentage fat 1.3-4.8% higher than UWW. In addition, the DEXA technique has the unique ability to produce precise regional measurements (for each arm, each leg, the head and the trunk) of bone mineral content (BMC), fat mass (FAT), lean mass (LEAN) and percentage fat.