Original articles   

The Journal of Sports Medicine and Physical Fitness 2000 June;40(2):178-83

Copyright © 2002 EDIZIONI MINERVA MEDICA

language: English

Body composition measurement in highly trained male athletes. A comparison of three methods

De Lorenzo A. ^{1, 2}, Bertini I. ¹, Iacopino L. ^{1, 2}, Pagliato E. ³, Testolin C. ³, Testolin G. ³

From the 1 Human Physiology University of Rome “Tor Vergata” 2 Scientific Institute “S. Lucia”, Rome 3 International Center for the Assessment of Body Composition University of Milan, Italy

Background. The pur­pose of the ­present ­study was to eval­u­ate the dif­fer­enc­es in the ­body com­po­si­tion esti­mate of high­ly ­trained ­male ath­letes by ­three dif­fer­ent meth­ods: ­dual-ener­gy X-ray absorp­tiom­e­try (DXA), bio­electri­cal impe­dance anal­y­sis (BIA) and skin­fold thick­ness meas­ure­ment (SFT).
Methods. The ­study ­occurred dur­ing a non-inten­sive train­ing peri­od. The meas­ure­ments ­were per­formed at the Human Physiology labor­a­to­ry. Participants: forty-­three ­male ath­letes (19 water­po­lo, 9 ­judo, 15 ­karate) who exer­cised reg­u­lar­ly at ­least ­three ­hours per day, six ­days per ­week. Measurements: DXA, BIA and SFT meas­ure­ments of per­cent­age fat ­mass (%FM) and the ­amount of fat-­free ­mass (FFM) ­were per­formed. Statistical anal­y­ses: the dif­fer­ent %FM esti­mates by the ­three meth­ods ­were com­pared ­using a one-way ANO­VA, ­with ­posthoc Bonferroni ­test ­when val­ues ­were sig­nif­i­cant­ly dif­fer­ent. The ­same ­test was ­used to com­pare FFM val­ues. The coef­fi­cient of vari­a­tion (CV%=100×SD/­mean) was cal­cu­lat­ed for ­each meth­ods. Methodological dif­fer­enc­es ­among the var­i­ous meth­ods ­were ana­lysed ­with the meth­od of Bland-Altman.
Results. ANO­VA ­test ­showed sig­nif­i­cant dif­fer­enc­es ­among the ­three meth­ods. In par­tic­u­lar, DXA sig­nif­i­cant­ly (p<0.001) over­es­ti­mat­ed %FM (12.4±4.1%) and under­es­ti­mate FFM (67.0±6.9 kg) ­respect to SFT (7.8±0.9% and 71.0±8.2 kg) and BIA (9.9±1.4% and 69.4±7.9 kg). Also Bland-Altman com­par­i­son ­among the esti­mates indi­cat­ed ­wide dif­fer­enc­es ­between meth­ods.
Conclusions. DXA pro­vides dif­fer­ent ­body com­po­si­tion esti­mates ­than ­those ­derived ­from skin­fold thick­ness meas­ure­ment and BIA, so ­that the meth­ods ­should not be ­used inter­change­ably.