The validity of commonly used adipose tissue body composition equations relative to dual energy X-ray absorptiometry (DXA) in Gaelic games players

Dual-energy X-ray absorptiometry (DXA) and adipose tissue percentage estimates (AT%) derived from regression based skinfold equations were compared. 35 Gaelic games players [20.9±1.7 years; 78.1±8.6 kg; 179.5±5.7 cm] underwent whole body fan beam DXA scans following a standardised protocol and assessment of skinfold thickness at 8 sites. Adipose tissue% from the sum of skinfolds and/or via body density were calculated for general and athlete specific equations (SKf-AT %). The relationship, i. e., proportional bias, fixed bias and random error (SEE) between DXA-AT% and AT% derived from the 6 skinfold equations were determined using least squares regression analysis. Skinfold AT% estimates were underestimated relative to DXA-AT% across all skinfold equations except that of Durnin and Wormersley [9] (D&W-Σ_4AT%) (16.7±3.4 vs. 16.6±4.0%). All equations demonstrated 95% prediction intervals ranges exceeding ~10%. Each equation failed to predict AT% relative to DXA within an accepted±3.5% anthropometric error rate. It is recommended that the conversion of absolute skinfold thickness to an AT% is avoided and that the skinfold equations assessed herein are not utilised in Gaelic games players. Alternate 'sum of skinfold' approaches should be considered.