Background: To evaluate the accuracy of body fat percentage (BF%) estimates from a portable, non -traumatizing ultrasound device compared to dual-energy X-ray absorptiometry (DXA) the reference technique.
Subjects: A total of 63 healthy men aged from 18 to 60 years participated. The patients were recruited according to a wide range of body mass index (BMI) and consequently of total fat mass. We also selected 100 athletes aged from 18 to 30 years from the French National Institute of Sports and Physical Education: 16 boxers, 4 rowers, 5 gymnasts, 6 base ball players, 19 judo players, 10 taekwondo, 7 basketball players, 21 wrestlers, 6 cyclists on track and 6 karate expert.
Methods: Ultrasound measurement (UT) were made with a sonographic US BOX in A-mode from Lecoeur Electronic Co.(Chuelles, France). We selected two preferred anatomical areas: the intra abdominal area which is often associated with metabolic risk factor and the mid- thigh area. Subcutaneous fat was located in a horizontal plane with approximately 45° axis vertebral at umbilical level and at the middle of the knee and the top thigh on anterior side.
BF(kg) and BF% measurements were obtained using a Hologic QDR-4500W (version 11.25; Hologic Bedford, Mass. USA) for the total sample. The DXA technique, which scans the whole body with an X-ray beam at two energy levels (70 and 100 Kev) is a reference method for measurement of fat mass, lean mass and mineral content. e subject lies in a supine position for 7 minutes and radiation exposure is very low.
The multiple linear regression to produce BF (kg) estimate with ultrasound dimensions on sedentary male is:
BF(kg) estimate = 0.708 BMI + 0.259 WC + 0.108 UTumb – 31.7

R² = 0.97, SEE = 1.7 kg
with BMI = Weight/Height (kg/m²) UTumb: umbilical thickness of fat (left + right)/2 side (mm); WC: waist circumference (cm).
The simplest model including BMI explained 90% of the variation of BF. Adding BMI and waist circumference in the model significantly increased the R² from 90 to 94% and decreased

SEE from 3.1 to 2.4. Moreover addition of ultrasonic measurement at umbilical level significantly increased the R² from 0.94 to 0.97 (P<0.01). The multiple linear regression equation to produce BF estimate with anthropometry and ultrasonic measurements on male athletes is: BF (kg) estimate = -4.99 BMI +0.109 BMI² +0.154 WC +0.627 UTmid-thigh + 46.19 with a concordance correlation of c =0.931. Pearson correlation R²=0.87 and SEE=1.6 kg. Predict BF% = 100 BF(kg) estimated ./ Body mass (kg). A cross-validation study was then performed with this linear regression on 62 male athletes proportionally stratified across the sports. The concordance correlation c =0.931 and SEE=1.61 kg. Conclusions: The use of an ultrasound portable device associated with anthropometric measurements allowed us to estimate %BF with a high level of accuracy according to the reference DXA. This ultrasonic technique can be used in sedentary subjects as well as high level sportsmen.