Author(s): VAN DIJK, J.1, BRINKMANS, N.1, PLASQUI, G.2, VAN LOON, L.2, Institution: HAN UNIVERSITY OF APPLIED SCIENCES, Country: NETHERLANDS, Abstract-ID: 1472

INTRODUCTION:
Accurate and practical assessment of body composition is essential for monitoring health and performance in professional football. Dual-energy X-ray Absorptiometry (DXA) provides whole-body and regional body composition, while deuterium dilution (DD) provides whole-body estimates of fat-free mass and fat mass. However, routine use of laboratory methods is often limited by cost and logistics, leading practitioners to rely on anthropometric assessments. This study compared laboratory-based and anthropometric methods for assessing whole-body and regional body composition in elite male and female footballers, to inform interpretation and use for routine monitoring in professional football.
METHODS:
Body composition was assessed in 58 male and 23 female Dutch Eredivisie players using DXA (Classic and NHANES algorithms), DD, and anthropometry following International Society for the Advancement of Kinanthropometry (ISAK) standardisation, including raw skinfolds, skinfold-based prediction equations, and girth measurements. Bland–Altman analysis assessed agreement between BF% estimates (DXA Classic, DXA NHANES, DD, and skinfold-based prediction equations), while Pearson or Spearman correlations examined associations between measures with different units (raw skinfolds and girths vs DXA-derived body composition).
RESULTS:
DXA Classic showed good agreement with DD for BF%, with a small mean difference (−0.64%; limits of agreement: −3.42 to 2.15%), whereas DXA NHANES systematically overestimated BF% (mean difference: 3.77%; limits of agreement: 0.91 to 6.64%). Skinfold-based prediction equations consistently underestimated BF% relative to DXA, with mean differences up to ~5% in males and up to ~10% in females, with wide limits of agreement, indicating large individual error. Raw anthropometric measures, however, showed strong associations with laboratory-derived body composition. Specifically, the sum of eight skinfolds showed strong associations with DXA-derived BF% (r = 0.81 in males, r = 0.88 in females) and moderate to strong associations with DD BF% (r = 0.63 in males, r = 0.81 in females), outperforming the sum of four skinfolds. Analyses of the trunk, arms, and legs demonstrated moderate to strong associations between regional skinfolds and DXA-derived regional fat mass (r = 0.55–0.89), while girths were moderately to strongly associated with regional lean mass (r = 0.58–0.85).
CONCLUSION:
DXA Classic shows good agreement with deuterium dilution for BF% assessment and is therefore preferred over DXA NHANES, which systematically overestimates BF%. Skinfold-based prediction equations show limited agreement with laboratory-derived BF% due to large individual bias. In contrast, raw anthropometric measures, including skinfold thicknesses and girths, adequately reflect whole-body and regional body fat and lean mass relative to laboratory measures, supporting their use for individual-level monitoring in elite male and female footballers.