This study compared %BFUS to %BF4C in young adult athletes. University club sport athletes (86 women, 138 men) from a variety of teams participated. ADP, DXA, and bioimpedance spectroscopy were used to measure body volume, bone mineral content, and total body water, respectively, for the 4C model. Ultrasound measurements were taken at three sites to estimate %BFUS. Comparisons between %BFUS and %BF4C were evaluated with Pearson correlation, paired t-test, linear regression, equivalence testing, and plots of individual errors. For the combined sample of men and women (N = 224), the two methods were highly correlated (r = 0.93) with a constant error (CE) of 0.8% BF (P < .001) and standard error of estimate (SEE) of 3.0% BF. The 90% confidence intervals of the mean difference (0.47 to 1.17) were well within the ±2% BF limits for equivalence testing, and no bias was evident from the error plot. However, equivalence testing failed to remain inside the ±2% BF range for women, and the prediction errors for women (CE = 1.9% BF, P < .001, SEE = 3.2% BF) were larger than for men (CE = 0.1% BF, P = .597, SEE = 2.8% BF). Overall, the %BFUS errors were low, suggesting that A-mode ultrasound is a valid field measure of %BF for young adult athletes. Accuracy is better for men than women.
Read full abstract