Obesity and serum lipids: an evaluation of the relative contribution of body fat and fat distribution to lipid levels
Foster, C.J.; Weinsier, R.L.; Birch, R.; Norris, D.J.; Bernstein, R.S.; Wang, J.; Pierson, R.N.; Van Itallie, T.B.
International Journal of Obesity 11(2): 151-161
1987
ISSN/ISBN: 0307-0565 PMID: 3610468 Document Number: 294113
The association of obesity and hyperlipidemia does not mean that fatness per se is the primary determinant of the lipid abnormality. To evaluate the contribution of fatness to fasting levels of serum triglycerides (TG), LDL cholesterol (LDL-C), and HDL cholesterol (HDL-C), we analyzed data on 368 caucasian adults (286 women, 82 men) consecutively entering a weight control program. Although most subjects were overweight, the population represented a wide spectrum of body weights and lipid levels. Study variables included body fat mass (by total body water), fat free mass (FFM), body build (chest to height ratio), fat cell size and number (from bilateral buttock biopsy specimens), upper-lower body fat pattern by arm to thigh circumference ratio, central-peripheral fat pattern by subscapular to triceps skinfold ratio, waist to hip ratio, and the presence or absence of diabetes. Our results concurred with previously noted correlations of body weight with TG (r = 0.29, P less than 0.0001) and with HDL-C (r = -0.28, P less than 0.0001) at least in the larger sample of women, but there was no significant correlation with LDL-C (r = -0.06). In order to evaluate the relative contribution of the various components of body composition and fat distribution to lipid levels, stepwise regression analyses were performed on the subgroups of women and men. Among women: TG level was predicted best by FFM, upper body fat pattern, age, and diabetes (explaining 30 percent of TG variance); LDL-C level was predicted by age only (explaining 12 percent of variance); and HDL-C level was predicted by body build only (8 percent). Among men: TG level was predicted best by central and upper body fat patterns and diabetes (31 percent of variance); LDL-C and HDL-C levels were not significantly predicted by any of the 11 study variables. These results, obtained from cross-sectional analysis of a predominantly obese sample, suggest that lipid levels may be more directly related to body fat pattern, fat free mass and body build than to body fatness itself.(author)