Variation in the anthropometric dimensions for estimating upper and lower body obesity

A literature review identified 29 anthropometric methods to estimate body fat distribution, but no valid standardized set of criterion values to categorize upper and lower body obesity. The purpose of this investigation was to observe the influence of different anthropometric methods for categorizing upper and lower body obesity. Three common girth methods (waist:iliac crest, iliac crest: hip, and waist:hip) for waist to hip ratios (WHR) and one common skinfold method for trunk to extremity ratio were compared in 119 obese women (38.4 ± 5.3% fat) and 81 obese men (31.3 ± 5.2% fat). Significant differences were found in the WHR among all three girth methods for women, but for only the waist:hip method for men. A significant interaction between method and sex was also apparent. The different WHR methods resulted in a significantly different percentages of the population categorized as having lower, normal, and upper body obesity. Kappa statistics demonstrated significant agreement among the three methods to place subjects into similar fat distribution categories for men (κ = 0.133; z = 4.03), but not for women (κ = 0.059; z = 2.20). There was also no significant correlation between any of the circumference and skinfold methods for men (r = 0.20, 0.10 and 0.15, respectively); yet, significant correlations were found between the iliac crest:hip (r = 0.30) and waist:hip (r = 0.45) methods and the skinfold method for women. In summary, girth methods for estimating body fat distribution in terms of lower and upper body obesity need validation and standardization. © 1994 Wiley-Liss, Inc.