(29) USE OF WAIST-TO-HEIGHT AND HIP-TO-HEIGHT RATIOS TO ESTIMATE % FAT IN COLLEGE AGE MEN AND WOMEN

Abstract Details: Recent focus on circumference-to-height ratios has renewed discussion of the use of
anthropometry to estimate %fat and ideal body weight (IBW). Considering the implications of
body composition for military service, it seems prudent to evaluate the feasibility of waist (WC)
and hip (HC) circumferences relative to height to estimate %fat in military eligible personnel.

Purpose: To evaluate the relationships of WC/Ht, HC/Ht, and BMI to %fat in young college
men and women.

Methods: Men (n = 1501) and women (n = 1282) volunteered to be
measured for 3-skinfold estimate of %fat. WC was measured at the umbilicus and HC was
determined at maximum gluteal protrusion. WC/Ht and HC/Ht were calculated by dividing each
circumference (cm) by height (cm).

Results: In men, WC/Ht (r = 0.75), HC/Ht (r = 0.70),
and BMI (r = 0.73) were significantly correlated with %fat. In women, HC/Ht (r = 0.71), WC/Ht
(r = 0.66), and BMI (r = 0.74) were also significantly correlated with %fat. Regression analysis
to predict %fat selected sex and HC/Ht in that order as primary variables in a validation group (n
= 1267) (R = 0.87, SEE = 4.2%). Sex made a slightly larger percent contribution to the equation
(51%) than did HC/Ht (49%). When the sexes were evaluated separately, regression analysis in
men selected BMI and WC.Ht as primary predictors of %fat (R = 0.77, SEE = 3.7%). In women,
BMI and HC/Ht were the primary predictors of %fat (R = 0.76, SEE = 4.0%). Cross-validation
of the men’s equation on a randomly selected sample (n = 408) revealed predicted %fat (11.0 ±
4.1%) was not significantly different (p = 0.22) from actual %fat (10.8 ± 5.4%), and they were
significantly correlated (r = 0.75). In women (n = 319), there was also no significant difference
(p = 0.39) between predicted (23.4 ± 5.7%) and actual %fat (23.2 ± 4.5%) with a significant
correlation between the two (r = 0.72).

Conclusion: %fat can be predicted in military-age
men and women using a combination of BMI and sex-specific variables of WC/Ht or HC/Ht
with an acceptable level of accuracy. Sex-specific equations used different circumference-to-
height ratios to estimate %fat with equal efficiency.

Acknowledgements: None