Skip to main content
ARS Home » Plains Area » Houston, Texas » Children's Nutrition Research Center » Research » Publications at this Location » Publication #205477

Title: Cross-validation and calibration of Jackson-Pollock equations with DXA: the TIGER study

Author
item JACKSON, A - UNIV OF HOUSTON DEP HPP
item ELLIS, K - USDA/ARS CNRC
item BUSH, J - UNIV OF HOUSTON DEP HPP
item TURPIN, IAN - USDA/ARS CNRC
item MILLER, F - UNIV OF HOUSTON DEP HPP
item CALLIE, M - USDA/ARS CNRC
item Bray, Molly

Submitted to: Medicine and Science in Sports and Exercise
Publication Type: Abstract Only
Publication Acceptance Date: 9/1/2005
Publication Date: 10/5/2005
Citation: Jackson, A.S., Ellis, K.J., Bush, J., Turpin, I., Miller, F., Callie, M.E., Bray, M.S. 2005. Cross-validation and calibration of Jackson-Pollock equations with DXA: the TIGER study [abstract]. Medicine and Science in Sports and Exercise. 37(5 Suppl):S302.

Interpretive Summary:

Technical Abstract: The Jackson-Pollock (J&P) body composition equations were developed primarily from data on white men and women using hydrostatically determined body density (BD) as the criterion measure. This study cross-validated the J&P equations with ethnically diverse subjects and percent fat (%fat) determined by dual energy X-ray absorptiometry (DXA). The sample included 109 females and 49 males (mean age = 28.1 +/- 10.6) consisting of the following ethnic groups: white, 35.4%; African-American, 32.9%; Hispanic, 17.1%; and other, 14.6%. The male and female DXA %fat means were 24.4 +/- 7.1 and 36.3 +/- 7.1. The J&P equations were used to estimate BD from age and sum of three skinfolds. The Siri equation converted estimated BD to %fat. The correlation between measured and estimated %fat was 0.92. The J&P %fat underestimated DXA %fat by 0.9% (+/- 3.95). A Bland-Altman plot showed a significant (p = 0.013) inverse relationship between DXA and JP %fat residuals and means. Regression analysis was used to calibrate J&P to DXA %fat. The model's slope and intercept were 0.85 and 5.76, respectively (R = 0.92, SEE = 3.51). Ethnicity added to the model as a nominal variable did not account for additional DXA %fat variance. These results showed that the validity of the J&P equations were robust with DXA %fat measured on ethnically diverse subjects.