PURPOSE: To examine the intermethods agreement of dual-energy X-ray absorptiometry (DXA) and foot-to-foot bioelectrical impedance analysis (BIA) to assess the percentage of body fat (%BF) in young male athletes using air-displacement plethysmography (ADP) as the reference method. METHODS: Standard measurement protocols were carried out in 104 athletes (40 swimmers, 37 footballers, and 27 cyclists, aged 12-14 y). RESULTS: Age-adjusted %BF ADP and %BF BIA were significantly higher in swimmers than footballers. ADP correlates better with DXA than with BIA (r = .84 vs r = .60, P < .001). %BF was lower when measured by DXA and BIA than ADP (P < .001), and the bias was higher when comparing ADP versus BIA than ADP versus DXA. The intraclass correlation coefficients between DXA and ADP showed a good to excellent agreement (r = .67-.79), though it was poor when BIA was compared with ADP (r = .26-.49). The ranges of agreement were wider when comparing BIA with ADP than DXA with ADP. CONCLUSION: DXA and BIA seem to underestimate %BF in young male athletes compared with ADP. Furthermore, the bias significantly increases with %BF in the BIA measurements. At the individual level, BIA and DXA do not seem to predict %BF precisely compared with ADP in young athletic populations.
PURPOSE: To examine the intermethods agreement of dual-energy X-ray absorptiometry (DXA) and foot-to-foot bioelectrical impedance analysis (BIA) to assess the percentage of body fat (%BF) in young male athletes using air-displacement plethysmography (ADP) as the reference method. METHODS: Standard measurement protocols were carried out in 104 athletes (40 swimmers, 37 footballers, and 27 cyclists, aged 12-14 y). RESULTS: Age-adjusted %BF ADP and %BF BIA were significantly higher in swimmers than footballers. ADP correlates better with DXA than with BIA (r = .84 vs r = .60, P < .001). %BF was lower when measured by DXA and BIA than ADP (P < .001), and the bias was higher when comparing ADP versus BIA than ADP versus DXA. The intraclass correlation coefficients between DXA and ADP showed a good to excellent agreement (r = .67-.79), though it was poor when BIA was compared with ADP (r = .26-.49). The ranges of agreement were wider when comparing BIA with ADP than DXA with ADP. CONCLUSION: DXA and BIA seem to underestimate %BF in young male athletes compared with ADP. Furthermore, the bias significantly increases with %BF in the BIA measurements. At the individual level, BIA and DXA do not seem to predict %BF precisely compared with ADP in young athletic populations.
Authors: Pia Christensen; Thomas Meinert Larsen; Margriet Westerterp-Plantenga; Ian Macdonald; J Alfredo Martinez; Svetoslav Handjiev; Sally Poppitt; Sylvia Hansen; Christian Ritz; Arne Astrup; Laura Pastor-Sanz; Finn Sandø-Pedersen; Kirsi H Pietiläinen; Jouko Sundvall; Mathijs Drummen; Moira A Taylor; Santiago Navas-Carretero; Teodora Handjieva-Darlenska; Shannon Brodie; Marta P Silvestre; Maija Huttunen-Lenz; Jennie Brand-Miller; Mikael Fogelholm; Anne Raben Journal: Diabetes Obes Metab Date: 2018-08-07 Impact factor: 6.577