PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Obesity (Silver Spring). Author manuscript; available in PMC 2011 August 16.
Published in final edited form as:
PMCID: PMC3156571
NIHMSID: NIHMS311659

Sex Differences in the Association of Thigh Fat and Metabolic Risk in Older Adults

Abstract

We have previously shown a favorable association of subcutaneous leg fat with markers of insulin resistance and dyslipidemia in postmenopausal women. It is not known whether there is a sex dimorphism in the association of lower-body adiposity with reduced metabolic risk. Thus, our primary aim was to determine whether the favorable association of thigh subcutaneous fat, independent of abdominal fat, is also observed in older men. Mid-thigh and abdominal fat areas were measured by computed tomography (CT) in 108 older men and postmenopausal women (mean ± s.d.; 69 ± 7 years). Additionally, trunk and leg fat mass (FM) were measured by dual-energy X-ray absorptiometry (DXA). Markers of insulin resistance and dyslipidemia were determined from oral glucose tolerance tests and lipid and lipoprotein measurements, respectively. Outcomes were fasted and postchallenge (area under the curve, AUC) insulin (INSAUC) and glucose (GLUAUC), product of the insulin and glucose AUC (INSAUC × GLUAUC), triglycerides (TG), and high-density lipoprotein (HDL)-cholesterol. Consistent with our previous findings in postmenopausal women, adjusting for DXA trunk FM revealed a favorable association of DXA leg FM with the metabolic risk outcomes in both older men and postmenopausal women. Likewise, adjusting for CT abdominal visceral fat generally revealed a favorable association of CT thigh fat with metabolic risk outcomes in women, but not men. The discordance between the DXA and CT results in men is unclear but may be due to sex differences in visceral fat accrual. The mechanisms underlying the protective effect of thigh fat on metabolic risk factors need to be elucidated.

Introduction

The sex dimorphism in cardiovascular disease risk is often attributed to sex differences in body fat distribution. The reduced risk of cardiovascular disease in women compared to men (at least before menopause) has been speculated to be related to a reduced abdominal fat accumulation in women (1). Although lowerbody fat has been considered relatively benign with respect to disease risk (24), we (5,6) and others (716) have demonstrated that lower-body adiposity may actually be protective, rather than simply less harmful than abdominal fat. We observed favorable associations of lower-body adiposity with markers of insulin resistance and dyslipidemia in postmenopausal women after adjusting for upper-body adiposity (5). Moreover, studies that included both older women and men have demonstrated favorable associations of leg subcutaneous fat, independent of abdominal fat, with glucose and lipid concentrations (6,14,16). Indeed, greater subcutaneous thigh adipose tissue was associated with a lower prevalence of the metabolic syndrome in obese men and overweight and obese women (17). Despite these observations, sex differences in lower-body adiposity suggest that women have an increased propensity to store fat in thigh subcutaneous adipose tissue and away from abdominal visceral or, possibly, nonadipose tissue depots (e.g., muscle, liver). If true, we would hypothesize stronger inverse associations of subcutaneous thigh fat mass (FM) with metabolic risk factors in women than in men. Moreover, men generally have more visceral fat than women and our previous data in women demonstrated that even small amounts of visceral fat can overwhelm many of the benefits of lower-body fat (6). Thus, we hypothesized that any incremental benefit of thigh fat on metabolic risk would be countered by the larger visceral FM in men.

The general aim of this study was to determine whether there are sex differences in the associations of subcutaneous thigh fat with markers of insulin resistance and dyslipidemia. We postulated that the favorable independent associations of leg fat with metabolic risk factors would be stronger in women than men, particularly after accounting for visceral adiposity.

Methods and Procedures

Subjects

We measured body composition and select disease risk factors in 108 healthy older men (n = 53) and postmenopausal women (n = 55). The data were obtained as a part of baseline assessments for a randomized, double-blinded, placebo-controlled trial of dehydroepiandrosterone replacement in older adults. Complete details for this study population were previously reported (18), but are described here briefly. Participants met the following eligibility criteria: serum dehydroepiandrosterones <140 μg/dl (3.8 μmol/l); fasted serum triglycerides (TG) <400 mg/dl (4.52 mmol/l); blood pressure <180/95 mm Hg; normal thyroid function; normal liver enzymes. Study participants were nonsmokers and had not used hormone therapy in the past year. All women were at least 1 year past menopause (mean ± s.d.; 22 ± 10 years). Most women (~84%) had used estrogen-based hormone therapy for an average duration of 7 ± 7 years, but had stopped hormone therapy an average of 6 ± 8 years before study entry. All of the participants provided written informed consent to participate in the study, which was approved by the Colorado Multiple institutional review board.

Body composition

Total and regional FM (trunk FM and leg FM) was determined by dual-energy X-ray absorptiometry (DXA) using either Lunar DPX-IQ (n = 74; 31 women, 43 men; software version 4.38; Lunar, Madison, WI) or Hologic Delphi-W (n = 34; 24 women, 10 men; software version 11.2; Hologic, Bedford, MA). The recommendations of the manufacturers were used to define the trunk and leg regions. Because the use of two DXA instruments could not be avoided, we did a separate study of 34 subjects (previously reported in ref. (19)) measured on both instruments. The between-instrument bias was assessed using Bland–Altman plots and an orthogonal regression of the measures from Lunar regressed on Hologic. The mean ± s.d. of the bias for body mass, FM, and fat-free mass were 0.21 ± 0.69, 0.71 ± 1.12, and −0.92 ± 1.06 kg, respectively. The estimated regression coefficients from the orthogonal regression were used to transform data for the 74 subjects measured using a Hologic instrument in the present study.

Abdominal (visceral and subcutaneous) and mid-thigh (subcutaneous and intermuscular) fat areas were determined by computed tomography (CT) using a General Electric (Waukesha, WI) high-speed CT. Single axial CT images (120 kVp, 200–300 mAs, and 10-mm slice thickness) were acquired at the levels of the L2–L3 and the L4–L5 intervertebral spaces and the mid-thigh (20 cm proximal to the distal border of the lateral condyle). Adipose tissue areas were determined using a CT intensity range from image-generated histograms of adipose and soft tissue regions. The abdominal visceral fat areas (VFA; cm2) were manually outlined by tracing the muscles of the abdominal wall. Abdominal subcutaneous fat areas (SFA; cm2) were calculated by subtracting the VFA from the total abdominal fat area. Thigh SFA was separated from intermuscular fat area by manually tracing along the deep fascial plane surrounding the muscles. Intermuscular fat areas (cm2) were calculated by subtracting the thigh SFA from the total thigh fat area. Fat areas were averaged over the two abdominal slices and the right and left thigh slices.

Glucose tolerance test

A 75-g oral glucose tolerance test was administered in the morning after an overnight fast. Blood samples were obtained before and 30, 60, 90, and 120 min after glucose ingestion for glucose and insulin determinations. The total areas under the glucose (GLUAUC) and insulin (INSAUC) curves were calculated using the trapezoidal rule. The INSAUC and fasted insulin (INS0) were used as indexes of hyperinsulinemia and the product of the insulin and glucose areas (INSAUC × GLUAUC) was calculated as an index of peripheral insulin resistance (20,21).

Hormones and Metabolites

Blood samples were stored at −80 °C and analyzed in batch by the Core Laboratory of the General Clinical Research Center. Serum insulin concentrations were determined with a double-antibody radioimmunoassay (Diagnostic Systems Laboratory, Webster, TX). Serum glucose was measured using a hexokinase assay on a Cobra Mira Plus instrument (Roche Diagnostic Systems, Indianapolis, IN). Intra- and interassay coefficients of variation were 5.2 and 9.8% for insulin measurements; and 1.1 and 3.6% for glucose determinations.

Blood lipids and lipoproteins

Measurements of serum lipid and lipoprotein concentrations were done by the General Clinical Research Center Core Laboratory. Total cholesterol, high-density lipoprotein (HDL)-cholesterol and TG were measured by automated enzymatic commercial kits on a Cobra Mira Plus instrument (Roche Diagnostic Systems). Intra- and interassay coefficients of variation were as follows: (i) total cholesterol, 5.1 and 2.4%; (ii) TG, 1.4 and 3.3%; (iii) HDL, 4.5 and 2.9%. Low-density lipoprotein (LDL)-cholesterol was calculated using the Friedewald equation (22).

Statistics

The primary outcome variables (INS0, GLU0, INSAUC, GLUAUC, INSAUC × GLUAUC, TG, HDL) for analysis were chosen a priori based on our previous observations (5,16). The distributions for all measures were examined graphically for symmetry; INSAUC, GLUAUC, INSAUC × GLUAUC, and TG were right skewed and were log transformed for analysis. Agreement among the DXA and CT regional body composition measures was assessed using Pearson's product-moment correlation coefficient. Fisher's z transformation was used to compare the correlations between women and men. The interaction between sex and leg fat (DXA leg or CT mid-thigh) for each of the outcome variables were examined graphically and statistically. Although none of the interaction terms reached significance at the 0.05 level, the graphs and statistical trends (P = 0.2–0.5) suggested sex-by-leg fat interactions (CT measures only) for many of the outcomes (example provided in Figure 1). Therefore, we conducted sex-specific analyses. Moreover, there were statistically significant differences between women and men for many of the correlations among the various depots (Table 1) that further supported separate regression analyses.

Figure 1
The residuals of the regressions for (a) leg fat mass (FM) or (b) thigh subcutaneous fat area (SFA) with one outcome (insulin resistance; INSAUC × GLUAUC) were plotted by sex to illustrate the similarity in slopes between women and men for the ...
Table 1
Interrelatedness of DXA and CT body composition measures by sex

Associations between the measures of body composition (independent variables) and measures of glucose, insulin, and lipids (dependent variables) were evaluated by linear regression. The body composition measures were added one at a time, thigh fat first, to illustrate the independent effect of each measure and how they behave in concert. Regression coefficients are reported as standardized betas (β), which capture the change in s.d. units of the dependent variable associated with a 1 s.d. change in the independent variable. The tolerance was not <0.2 for any independent variable, so we judged the models to be reasonably stable. All statistical analyses were performed using SAS version 9.1 (SAS Institute, Cary, NC).

Results

Subject Characteristics

Women and men were similarly overweight by BMI (27 kg/m2), but women had more FM (Table 2). Compared to men, women had 40% more subcutaneous abdominal fat, 107% more thigh subcutaneous fat, and 36% less visceral fat. Women also had higher total cholesterol than men due to higher LDL- and HDL-cholesterol levels. Medication use among women and men, respectively, was as follows: (i) antihypertensives (18 and 9%); (ii) lipid-lowering drugs (15 and 20%); (iii) thyroid replacement (20 and 9%); (iv) oral antidiabetic agents (2 and 2%), and (v) aspirin (31 and 49%). Twice as many women than men were on antihypertensive or thyroid medication.

Table 2
Subject characteristics by sex (mean ± s.d.)

Leg and trunk FM measured by DXA were highly correlated with thigh and abdominal fat areas measured by CT (Table 1). However, the interrelatedness of many of these variables differed significantly between women and men. In men, abdominal VFA was not correlated with leg FM or thigh SFA and was more weakly correlated with abdominal SFA. In women, abdominal VFA was highly correlated with all measures of body fat.

DXA-derived regional fat

Linear regression models (Table 3) were used to evaluate the associations of DXA leg FM with each of the primary outcomes before (model 1) and after adjusting for trunk FM (model 2) in women and men separately. In general, there was no association of leg FM with the outcomes in women or men (model 1). Only the associations with fasted insulin and HDL-cholesterol in women were statistically significant; none of the associations achieved significance in the men. After adjusting for the detrimental effects of trunk FM (model 2), leg FM became favorably associated with metabolic outcomes in both women and men. Although not all β coefficients reached statistical significance after adjustment (model 2), all β coefficients consistently changed from null or positive to negative with the exception of HDL which changed from negative to positive as expected. For example, the standardized β coefficient for the association of INS0 with leg FM changed from positive (0.344) to negative (−0.394) after adjusting for trunk FM in women; a similar reversal occurred in men (0.127 to −0.312). Thus, after adjusting for trunk FM, increased leg FM was associated with lower fasting insulin concentration in women and men. Similar patterns were observed for each metabolic risk outcome variables.

Table 3
Linear regression models for women (n = 55) and men (n = 53) separately for each dependent variable entering DXA leg fat mass (leg FM; model 1) and adjusted for DXA trunk fat mass (trunk FM; model 2)

CT-derived regional fat

Linear regression models (Table 4) were used to evaluate the associations of thigh SFA with each of the primary outcomes before (model 1) and after adjusting for abdominal VFA (model 2) in women and men separately. In general, the independent associations of thigh SFA with the metabolic outcomes were of similar magnitude and direction as the associations of DXA leg FM with the outcomes. Further, as in the DXA models, adjusting for the detrimental effects of visceral fat on metabolic risk outcomes resulted in a shift in the β coefficients for the association between thigh SFA and each of the metabolic outcomes in women. For example, the β coefficient for the association of INS0 with thigh SFA changed from positive (0.383; P < 0.05) to negative (−0.132; P = 0.34) afer adjusting for abdominal VFA. Although none of the adjusted β coefficients for thigh SFA reached statistical significance, with the exception of fasting glucose, there was a consistent shift from null or positive to negative (the reverse for HDL) for all outcomes in women, as was observed in the DXA models, but the same was not true in men. In men, the β coefficient for the association between thigh SFA and each of the outcomes did not change at all after adjustment for abdominal VFA. For example, in men there was no association between INS0 and thigh SFA (0.013) and this did not change after the adjustment for abdominal VFA (0.000). Thus, the CT models were not consistent with the DXA models for the men.

Table 4
Linear regression models for women (n = 55) and men (n = 53) separately for each dependent variable entering CT mid-thigh subcutaneous fat area (SFA; model 1) and adjusted for CT abdominal visceral fat area (VFA; model 2)

Discussion

The results of this study of older women and men confirmed our previous observations in postmenopausal women (5) that DXA leg FM was associated with reduced insulin resistance and dyslipidemia, independent of the increased risk attributable to DXA trunk FM. After adjusting for trunk FM, the associations (β coefficients) between leg FM and each of the metabolic outcomes were improved (e.g., positive association for HDL, negative association for INS, GLU, and TG). In the DXA models, the magnitudes of these associations were generally similar in women and men.

Because abdominal visceral fat, which is not distinguished from abdominal subcutaneous fat by DXA, is thought to confer the most metabolic risk, we also used CT-derived measures of adiposity to assess the independent associations of leg subcutaneous fat with metabolic risk outcomes. As in the DXA models in women, favorable associations of CT thigh subcutaneous fat with metabolic risk were generally observed after accounting for the unfavorable contribution of CT abdominal visceral fat, although most did not reach statistical significance. However, in men the opposing contributions of thigh and abdominal fat to metabolic risk that were apparent in the DXA models were not apparent in the CT models. This was likely due to the lack of association (r = 0.02) between thigh SFA and abdominal VFA in men.

Although overall adiposity is greater in premenopausal women than age-matched men, their cardiovascular disease risk is much less (4,6). The preponderance of abdominal visceral adipose tissue in men is thought to be an important contributor to their increased metabolic risk (2). Conversely, lower-body adiposity in premenopausal women has been postulated to be cardioprotective (1). This apparent cardioprotection appears to diminish after menopause when abdominal fat readily accumulates (23,24). The relative amount of fat stored in the lower body compared to the upper body may be important in disease progression. Indeed, the independent and opposing effects of waist and hip girth on myocardial infarction risk were demonstrated in the INTERHEART Study (25). After adjusting for age, sex, smoking, and BMI, this study of >27,000 women and men from 52 countries found an increasing incidence of myocardial infarction with increasing waist size (odds ratio 1.77, 1.59–1.97; top vs. bottom quintile; P < 0.0001), but a decreasing incidence of myocardial infarction with increasing hip size (odds ratio 0.73, 0.66–0.80; top vs. bottom quintile; P < 0.0001).

Although several studies have reported an inverse association of lower-body adiposity with various disease risk factors in populations that have included both women and men (8,1214), few have addressed whether a sex dimorphism exists. To our knowledge, Health ABC was the only study to evaluate this separately in older women and men, and there was a favorable association of subcutaneous thigh fat (measured by CT) with glucose and lipid concentrations, independent of abdominal fat, in both sexes (16). Whereas our DXA models corroborated these results, our CT models did not. In contrast to the Health ABC study, we did not observe a consistent association of CT thigh SFA with metabolic risk factors, independent of CT abdominal VFA, in men. It is not clear why our CT results differed. It is possible that, because of our smaller number of subjects than Health ABC and the low variability in thigh subcutaneous fat in men, our study was not adequately powered to detect significant contributions of thigh SFA independent of VFA in men. Indeed, it is important to note that we were underpowered to detect sex-by-leg fat interactions and to reach statistical significance in the adjusted β coefficients in many of our outcomes. However, it was the consistency in our results across all outcomes that gave us confidence in our overall conclusions. That is, there was a consistent shift in the directionality of the β coefficient for women and men in the DXA models and for women in the CT models which suggested a favorable effect of leg FM or thigh SFA on risk, independent of trunk FM or abdominal VFA. On the other hand, the CT data in men consistently suggested no beneficial effect of thigh SFA after adjusting for abdominal VFA. Further, given that there weren't even small directional changes in these adjusted β coefficients, we do not believe a larger sample size would have changed the CT results in men. Clearly, these results need confirmation in other settings; consistency across multiple studies will give greater confidence that the observed associations are real.

It is not clear why our DXA and CT models provided generally consistent results for women, but discordant results for men. It is likely that this is due to the sex-differences in the simple correlations between leg subcutaneous fat and visceral fat (Table 1). Leg subcutaneous fat (measured via DXA or CT) was strongly correlated with visceral fat in women (r = 0.68–0.69), but not in men (r = 0.02–0.17), suggesting that abdominal visceral fat accumulation is relatively independent of thigh subcutaneous fat accumulation in men. However, the possibility remains that greater leg fat accumulation is indicative of lesser visceral fat accumulation when these simple correlations are considered in the context of total body mass. Indeed, after controlling for BMI, leg subcutaneous fat was inversely related to abdominal visceral fat in men (r = −0.44, P = 0.001) but not in women (r = −0.16, P = 0.24). Thus, although women have a propensity to store fat in the lower body, more leg fat does not predict less visceral fat for a given BMI in women. On the other hand, although men have a propensity to store fat in the visceral compartment, increased leg fat does predict less visceral fat for a given BMI in men. Taken together these data suggest there may be sex-differences in the association between abdominal visceral and leg subcutaneous fat accumulation.

It could be argued that the difference in body fat between the women and men (40.5 vs. 28.6%) influenced the observed sex differences in our CT data. We know that women, on average, have a greater relative body fat content than men throughout the lifespan, so differences in total adiposity are a consideration whenever women and men are compared. However, if we had a cohort of women and men who were matched for relative adiposity (i.e., included fatter men and/or leaner women) the sex-differences in regional adiposity would remain and visceral adiposity would likely be exaggerated. Thus, given that abdominal visceral fat in men was not correlated with leg subcutaneous fat and did not change the association of leg fat with any of the outcomes when entered into the regression models (Table 4), we believe the CT results would remain unchanged in a cohort of women and men matched for relative adiposity.

In summary, our DXA data suggest that the independent favorable association of leg fat with risk factors for insulin resistance and dyslipidemia exists in both women and men despite sexrelated differences in the amount of fat stored in the legs. The DXA data suggest the metabolic benefit (or lack of harm) from storing fat in the lower body is the same for women and men. In contrast, our CT data suggest that there is little benefit of leg fat in men after visceral adiposity is taken into account.

Acknowledgments

The authors thank the staffs of the UCD General Clinical Research Center (GCRC) and Energy Balance Core of the Clinical Nutrition Research Unit (CNRU) for their assistance in conducting this study. The authors also like to thank the members of their research group for carrying out the day-to-day activities of the project and the study volunteers for their time and efforts. The following awards from the National Institutes of Health supported this research: R01 AG018198 and AG018857, K01 AG019630 (R.E.V.), T32 AG000279 (C.M.J.), F32 AG005899 (W.S.G.), M01 RR000051 (GCRC), P30 DK048520 (CNRU).

Footnotes

Disclosure: The authors declared no conflict of interest.

References

1. Kannel WB, Wilson PW. Risk factors that attenuate the female coronary disease advantage. Arch Intern Med. 1995;155:57–61. [PubMed]
2. Lemieux S, Després JP, Moorjani S, et al. Are gender differences in cardiovascular disease risk factors explained by the level of visceral adipose tissue? Diabetologia. 1994;37:757–764. [PubMed]
3. Vague J. La differenciation sexuelle, facteur determinant des formes do l'obesite. Presse Med. 1947;30:339–340. [PubMed]
4. Wingard DL, Suarez L, Barrett-Connor E. The sex differential in mortality from all causes and ischemic heart disease. Am J Epidemiol. 1983;117:165–172. [PubMed]
5. Van Pelt RE, Evans EM, Schechtman KB, Ehsani AA, Kohrt WM. Contributions of total and regional fat mass to risk for cardiovascular disease in older women. Am J Physiol Endocrinol Metab. 2002;282:E1023–E1028. [PubMed]
6. Van Pelt RE, Jankowski CM, Gozansky WS, Schwartz RS, Kohrt WM. Lower-body adiposity and metabolic protection in postmenopausal women. J Clin Endocrinol Metab. 2005;90:4573–4578. [PMC free article] [PubMed]
7. Williams MJ, Hunter GR, Kekes-Szabo T, Snyder S, Treuth MS. Regional fat distribution in women and risk of cardiovascular disease. Am J Clin Nutr. 1997;65:855–860. [PubMed]
8. Terry RB, Stefanick ML, Haskell WL, Wood PD. Contributions of regional adipose tissue depots to plasma lipoprotein concentrations in overweight men and women: possible protective effects of thigh fat. Metab Clin Exp. 1991;40:733–740. [PubMed]
9. Freedman DS, Rimm AA. The relation of body fat distribution, as assessed by six girth measurements, to diabetes mellitus in women. Am J Public Health. 1989;79:715–720. [PubMed]
10. Hunter GR, Kekes-Szabo T, Snyder SW, et al. Fat distribution, physical activity, and cardiovascular risk factors. Med Sci Sports Exerc. 1997;29:362–369. [PubMed]
11. Tankó LB, Bagger YZ, Alexandersen P, Larsen PJ, Christiansen C. Peripheral adiposity exhibits an independent dominant antiatherogenic effect in elderly women. Circulation. 2003;107:1626–1631. [PubMed]
12. Seidell JC, Pérusse L, Després JP, Bouchard C. Waist and hip circumferences have independent and opposite effects on cardiovascular disease risk factors: the Quebec Family Study. Am J Clin Nutr. 2001;74:315–321. [PubMed]
13. Sjöström CD, Håkangård AC, Lissner L, Sjöström L. Body compartment and subcutaneous adipose tissue distribution–risk factor patterns in obese subjects. Obes Res. 1995;3:9–22. [PubMed]
14. Snijder MB, Dekker JM, Visser M, et al. Trunk fat and leg fat have independent and opposite associations with fasting and postload glucose levels: the Hoorn study. Diabetes Care. 2004;27:372–377. [PubMed]
15. Okura T, Nakata Y, Yamabuki K, Tanaka K. Regional body composition changes exhibit opposing effects on coronary heart disease risk factors. Arterioscler Thromb Vasc Biol. 2004;24:923–929. [PubMed]
16. Snijder MB, Visser M, Dekker JM, et al. Low subcutaneous thigh fat is a risk factor for unfavourable glucose and lipid levels, independently of high abdominal fat. The Health ABC Study Diabetologia. 2005;48:301–308. [PubMed]
17. Goodpaster BH, Krishnaswami S, Harris TB, et al. Obesity, regional body fat distribution, and the metabolic syndrome in older men and women. Arch Intern Med. 2005;165:777–783. [PubMed]
18. Jankowski CM, Gozansky WS, Schwartz RS, et al. Effects of dehydroepiandrosterone replacement therapy on bone mineral density in older adults: a randomized, controlled trial. J Clin Endocrinol Metab. 2006;91:2986–2993. [PubMed]
19. Jankowski CM, Gozansky WS, Van Pelt RE, et al. Relative contributions of adiposity and muscularity to physical function in community-dwelling older adults. Obesity (Silver Spring) 2008;16:1039–1044. [PubMed]
20. Levine R, Haft DE. Carbohydrate homeostasis. I. N Engl J Med. 1970;283:175–183. [PubMed]
21. Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes Care. 1999;22:1462–1470. [PubMed]
22. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972;18:499–502. [PubMed]
23. Kotani K, Tokunaga K, Fujioka S, et al. Sexual dimorphism of age-related changes in whole-body fat distribution in the obese. Int J Obes Relat Metab Disord. 1994;18:207–202. [PubMed]
24. Pascot A, Lemieux S, Lemieux I, et al. Age-related increase in visceral adipose tissue and body fat and the metabolic risk profile of premenopausal women. Diabetes Care. 1999;22:1471–1478. [PubMed]
25. Yusuf S, Hawken S, Ounpuu S, et al. Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: a case-control study. Lancet. 2005;366:1640–1649. [PubMed]