Body composition assessment in athletes with spinal cord injury: comparison of field methods with dual-energy X-ray absorptiometry

Study design: Cross-sectional. Objectives: To compare relative body fatness (%Fat) estimates from field methods (skinfold thickness measurement (SKF) and bioelectrical impedance analysis (BIA)) with measures by dual-energy X-ray absorptiometry (DXA). Setting: University of Illinois, Urbana-Champaign, IL, USA. Methods: Field methods used both three-and seven-site SKF prediction equations and BIA generalized, spinal cord injury (SCI)-specific and athlete-specific equations. DXA was used as the reference method. College-aged varsity athletes with SCI (women = 8, men = 8; time since injury 16.2 +/- 5.7 years; injury level range T5-L5) were recruited. Results: Mean BMI was 20.8 +/- 2.6 and 22.5 +/- 2.1 kg m(-2), and mean DXA %Fat was 31.9 +/- 3.8 and 20.6 +/- 8.4%, for women and men, respectively. All field methods under-predicted the %Fat when compared with DXA (ranges in mean differences: SKF women 2.9-8.2%, SKF men 6.9-12.4%; BIA women 0.5-3.9%, BIA men 0.3-7.0%). None of the field methods accurately predicted the %Fat compared with DXA (total error (TE): SKF women 7.4-12.1%, SKF men 8.4-15.2%; BIA women 5.1-9.3%, BIA men 6.7-10.7%). Of the SKF and BIA prediction equations, Evans et al.'s three-site SKF (r = 0.95, P < 0.001, standard error of the estimate (SEE) = 2.8 %Fat) prediction equation provided the best fit for this population. Conclusion: Further studies with larger samples are necessary to develop appropriate prediction equations for field methods in the athletic SCI population. Spinal Cord (2009) 47, 698-704; doi:10.1038/sc.2009.20; published online 17 March 2009