Abstract des Autors

Accurate assessment of body composition is an important issue among athletes. Different methodologies generate controversial results, leading to a deep uncertainty on individual exercise prescriptions. Thus, this study aims to identify the differences between field methods, such as bioelectrical impedance (BIA) and skinfold assessment, with a clinical method, highly accurate, dual-energy X-ray absorptiometry (DXA), among elite young football players. Thirty-eight male football players with a mean (sd) age of 16.7 (0.87) years, involved in the Portuguese national competition of U16 (n = 13) and U19 (n = 25), were evaluated and objective measures of body composition, muscle strength and football skills were collected by trained specialists. Body composition was assessed using BIA (Tanita BC-418, Tanita Corp., Tokyo, Japan), in agreement with all the evaluation premises. Additionally, all athletes were evaluated using the clinical method DXA (Hologic Inc., Waltham, MA, USA). Among the U19 athletes, three skinfold sites (SKF) were assessed: chest, abdomin and thigh. The Spearman correlation coefficients and the mean difference between methods were calculated. The agreement between both methods was analyzed using Bland-Altman plots. Among the evaluated athletes, lower mean values of body fat % were found using BIA as a method of body composition assessment compared with DXA (12.05 vs. 15.58 for U16; 11.97 vs. 14.16 for U19). Despite the moderate correlation between methods (r = 0.33) to estimate the percentage of total fat, the median of the difference (DXA vs. BIA) was relevant in clinical terms, with 2.90% and 1.47% for U16 and U19 athletes, respectively. Stronger correlations were found between the sum of the SKF and DXA fat estimation (r = 0.68). The Bland-Altman plots showed a clear underestimation in the evaluations using the BIA, namely among athletes with better body composition profiles (8%–12% of fat). Using BIA, an underestimation of body fat assessment was observed among 94.5% of the athletes with less than 12% body fat mass. Among the evaluated athletes, fat mass was underestimated at a median value of 2.21% using BIA in comparison with DXA. The sum of the SKF showed a stronger correlation with the reference method (DXA) (r = 0.68) than BIA.