Total body fat estimations by near-infrared interactance, A-mode ultrasound, and underwater weighing

We studied 94 (41 men and 53 women) healthy 18 to 65 years old Caucasians for body composition utilizing 3 methods: 1) Underwater weighing (UWW) for body density, 2) A-mode ultrasound subcutaneous fat determination, and 3) Near-infrared (NIR) interactance spectral data. We then utilized the Wherry-Doolittle test selection method to predict percent of total body fat (%fat) from either ultrasound or NIR data with anthropometric measurements, optimizing for the fewest number of data points which correlated highly with UWW results. It was found that %fat could be estimated with a correlation of > 0.9 when measured at 4 sites (different between men and women) by ultrasound, age (women) and height (men). We also found a correlation of > 0.8 using 2 sites by NIR, weight, age, and height. It has been suggested that the ultrasound method would produce better estimates of both local and total body fat.