The measurement of body composition by infrared interactance: An evaluation

Abstract

The purpose of the present study was to investigate the validity of commercial instrumentation using near infrared interactance (NIRI) to predict total body fat. 169 adult men and women, aged 18–48 years, with markedly diverse body composition, were measured. The optical density (OD) at six sites (biceps, triceps, subscapular, suprailliac, thigh and calf) was measured by the NIRI technique with Futrex 5000. Subcutaneous skinfold thickness on thesame sites was measured by Harpenden caliper. Total body fat was estimated by densitometry and by Futrex manufacturer's equation. A good correlation between OD and skinfold thickness was obtained only on the biceps (r=0.81 for men and r=0.84 for women). The correlation between the percentage body fat (BF%) determined by densitometry and BF% predicted by the Futrex equation gave r=0.88 (SEE=4.0%) for men and 0.72 (SEE=4.7%) for women. The Futrex equation underestimated BF% (−2.1% in men and −3.9% in women), especially in fatter individuals. The predictive power of NIRI readings was tested by developing new prediction equations with the variables recommended by the manufacturer. The results whow that the presence of OD measures in the equation only slightly improves the predictive power of an equation based on weight, height, physical activity level and sex. From this study, it can be concluded that recourse to NIRI measures as obtained by currently available instrumentation, has no valid justification and that if the subject's body composition differs from the average normal, its use may introduce systematic biases and provide unreliable results.