Purpose: To cross-validate the DXA prediction of minimum weight (MW) in high school wrestlers, using a criterion-referenced analysis. The goal was to independently evaluate whether DXA provided a MW within acceptable limits for the sport of wrestling. Secondarily, the DXA prediction error was compared against the currently approved skinfold (SF) method.
Methods: Criterion MW was calculated by hydrostatic weighing (HW) with measured residual lung volume. Whole-body scans were performed with a Norland XR-36 bone densitometer. All skinfolds were taken by the same experienced measurer. The subject's body density was computed by Lohman and was converted to percent body fat, using the equation of Brozek et al. The measured fat-free mass was used to calculate each wrestler's MW at 7% body fat. Subjects were 94 Wisconsin high school wrestlers (mean (SD): age = 15.1 yr (1.2), height = 170.3 cm (7.1), weight = 63.2 kg (9.6).
Results: There was no significant difference in mean MW from DXA (60.6 kg (9.0)) and the HW criterion (59.8 kg (9.0)). The correlation was strong (r = 0.98), and the regression for the relationship between HW and DXA (y = 0.976 × DXA + 0.698 kg) did not significantly deviate from the line of identity. A low standard error of estimate (SEE) of 1.7 kg and a pure error (PE) of 1.9 kg were found, with residuals ranging from −3.94 to 2.88 kg. This PE was similar to the SF method (2.1 kg) in the sample. Bland-Altman analysis showed no systematic bias in the prediction of MW across weight classes.
Conclusion: We conclude that DXA provided a valid prediction of MW in this sample of high school wrestlers.
1University of Wisconsin Hospital Sports Medicine Center, University of Wisconsin, Madison, WI; 2Center for Childhood Obesity Research, The Pennsylvania State University, University Park, PA; and 3University of Wisconsin School of Medicine and Public Health, Department of Pediatrics, University of Wisconsin, Madison, WI
Address for correspondence: R. Randall Clark, University of Wisconsin Hospital Sports Medicine Center, University of Wisconsin Research Park, 621 Science Drive, Madison, WI 53711; E-mail: firstname.lastname@example.org.
Submitted for publication January 2007.
Accepted for publication July 2007.