Agreement Between 2 Segmental Bioimpedance Devices, BOD POD, and DXA in Obese Adults

This study examined the agreement between 2 segmental bioimpedance analysis (BIA) devices, air displacement plethysmography (BOD POD), and dual energy X-ray absorptiometry (DXA) for estimating body composition in obese adults. Fifty obese adults (25 men and 25 women; age = 34.20 +/- 11.19 years; BMI = 36.14 +/- 5.33 kg/m(2)) had their body fat percentage (BF%) and fat-free mass (FFM) evaluated with 2 segmental BIA devices (InBody 230 and InBody 720), BOD POD, and DXA (Lunar iDXA). Body composition via the BOD POD was determined using the Siri equation whereas manufacturer-based equations generated metrics (ie, BF% and FFM) for the InBody devices. The effect size of the mean differences for all BF% and FFM comparisons were trivial (Cohen's d < 0.20). The standard error of estimate (SEE), total error (TE), and 95% limits of agreement (LOAs) were low for both segmental BIA devices when compared to DXA (SEE < 2.26% and 2.35 kg; TE < 2.58% and 2.66 kg; 95% LOAs < +/- 4.94% and 4.86kg). The error for BOD POD was also low when compared to DXA (SEE = 2.39% and 2.57 kg; TE = 2.34% and 2.56 kg; 95% LOAs = 4.63% and 5.06 kg). Validity statistics were slightly higher, but considered acceptable, when comparing the segmental BIA devices against BOD POD (SEE < 3.37% and 3.63 kg; TE < 3.44% and 3.79 kg; 95% LOAs < +/- 6.62% and 7.19 kg). Lastly, the 2 segmental BIA devices produced nearly identical validity statistics when compared to each other. However, both BIA devices revealed proportional bias for BF% and FFM when compared to the BOD POD and DXA (all p < 0.05). The current study's findings indicate the InBody 230 is interchangeable with the InBody 720 in obese adults. Also, the trivial effect size, when compared against the BOD POD and DXA, suggest the InBody devices could be used for estimating group BF% and FFM. In contrast, the significant proportional bias demonstrates the BIA devices are not acceptable for individual estimates of body composition in an obese clinical population.