Near infrared reactance for the estimation of body fatness in regularly exercising individuals

Near infrared reactance (NIR) is used to measure body fat percentage (BF%), but there is little data on its use in non-obese, regularly exercising individuals. Therefore, this study aimed to examine the limits of agreement between NIR compared to dual x-ray absorptiometry (DXA) for the measurement of BF% in 2 cohorts of regularly exercising individuals. BF% was measured using DXA and NIR in a regular exercising (≥3 sessions/week), healthy active cohort (HA; n=57), and in a regularly exercising and resistance trained (≥2 sessions/week) cohort (RT; n=59). The RT cohort had lower BF% than the HA cohort (15.3±5.5% and 25.8±7.1%, P<0.001). In the HA and RT cohorts, NIR BF% was associated with DXA BF% (R2=0.72, SEE=3.7, p<0.001 and R2=0.50, SEE=4.1 p<0.001, respectively). In the HA cohort, NIR tended to under-predict BF% (mean difference: - 1.3%; 95% limits of agreement (LOA); - 8.8 to 6.2%) whereas in the RT cohort, NIR tended to over-predict BF% compared to DXA (mean difference: 1.1; 95% LOA; - 8.1 to 10.3%). In conclusion, NIR and DXA yield similar average BF% measurements in 2 cohorts of non-obese regularly exercising individuals. However, the rather broad LOA of NIR need to be considered when using NIR to screen for overweight and obesity, or measure and track changes in body composition.