Development Of Three Anthropometric Methods To Predict DXA-derived Body Fat Mass In Mexican Athletes

The measurement of body fat mass (BFM) is a fundamental component of fitness. Traditionally, the skinfold thickness (SKF) technique has been used for this purpose; however, its limitations in terms of precision and reliability are significant, leading to a quest to develop more robust indicators. In this context, using body surface area (BSA), central circumferences, and height might help improve BFM estimation from SKF. PURPOSE: To develop and validate BFM prediction models from anthropometric methods based on BSA, SKF, central circumferences, and height. METHODS: A total of 280 male athletes (weight: 72.09 ± 9.83 kg, height: 1.75 ± 0.08 m, age 23.9 ± 4.4; BFM: 11.9 ± 4.2 kg) underwent a whole-body DXA scan (Hologic QDR-4500) to assess their BFM as the reference. Anthropometric measurements were taken by certified personnel according to the ISAK protocol. The anthropometric variables included four central circumferences (chest, waist, abdomen, hips) and eight skinfolds (triceps, biceps, subscapular, iliac crest, supraspinale, abdominal, thigh, calf). We calculated three anthropometric methods to estimate BFM individually: Body Surface Area method (BSAM), Trunk Adiposity method (TAM), and Central Adiposity method (CAM). BSAM was calculated as = BSA * Average SKF (cm); the BSA was calculated as = (weight * height)/36000.5 TAM was calculated as = Height (m) * Average SKF (cm) * Average of Central circumferences (cm). CAM was calculated as = Height (m) * Average SKF (cm) * Average (Chest and Waist circumferences, cm). We used simple linear regressions to predict BFM from each anthropometric method and Bland-Altman analyses to compare the predicted scores with DXA. RESULTS: The mean BSAM, TAM, and CAM were 1.98 ± 0.78, 157.97 ± 62.39 and 164.09 ± 66.16, respectively. The obtained regression equations were BFM = 4.91 + 2.16 * BSAM (SEE = 1.82, R2 = 0.78), BFM =0.05 + 2.88 * TAM (SEE = 1.27 kg, R2 = 0.87) and BFM = 0.06 + 2.42 * CAM (SEE = 1.80, R2 = 0.79). The mean difference vs DXA (95% limits of agreement) was 0.10 (-3.79 to 3.99) for BSAM, 0.19 (-2.38 to 2.76) for TAM and 0.10 (-3.65 to 3.85) for CAM. CONCLUSIONS: The BSAM, TAM, and CAM are valid predictors for estimating BFM, and TAM seems to be the most accurate of them.