SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

Body fat distribution modulates risk for type 2 diabetes mellitus. We evaluated potentially involved metabolic risk factors. In a population of 282 male and 157 female healthy subjects (data from the San Antonio and the European Group of Insulin Resistance (EGIR) study cohorts), we evaluated associations between body fat distribution (assessed by waist and hip circumference) and parameters of lipid- and glucose metabolism, including clamp measurements of insulin sensitivity. After stratification for BMI, fasting triglycerides were lower in the presence of a large hip, and higher in a large waist. Persons with the largest BMI (3rd tertile) showed a difference in triglyceride levels of 1.5 vs. 2.4 mmol/l in large vs. small hip circumference groups (P < 0.038), and a difference of 1.5 vs. 1.2 mmol/l in large vs. small waist circumference groups (P < 0.025). A similar analysis did not reveal a difference in insulin sensitivity. Linear regression analyses confirmed the findings; they revealed negative associations between triglycerides and hip, and (for women borderline statistically significant) positive associations between triglycerides and waist, after adjustment for BMI, mutual confounding, and age (β ± s.e.; men: −0.48 ± 0.005, P < 0.001, and 0.21 ± 0.005, P < 0.05; women: −0.78 ± 0.007, P < 0.001, and 0.24 ± 0.005, P < 0.065), respectively. Linear regression analyses revealed similar opposite associations with high-density lipoprotein (HDL)-cholesterol, though not with glucose, insulin, or insulin sensitivity as measured with the clamp method. In our study population of healthy persons, body fat distribution is associated with fasting triglycerides and HDL-cholesterol, and not with insulin sensitivity. Metabolic risk of unfavorable body fat distribution may be modulated by lower triglyceride storage capacity.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

Obesity increases risk for cardiovascular disease and diabetes mellitus, but the increase in risk is not linear with the BMI. So called android obesity (apple shape) has a remarkably different risk than gynoid obesity (pear shape). Apparently, distribution of fat modulates risk. Abdominal obesity, clinically assessed by the “waist circumference,” is strongly associated with cardiovascular disease and diabetes mellitus (1,2,3), whereas “hip circumference” appeared to be protective (4). The waist circumference is regarded as the central feature of the metabolic syndrome (International Diabetes Federation 2005 criteria), which together with hypertension, lipid disturbances (increased triglycerides and decreased high-density lipoprotein (HDL)-cholesterol), and impaired fasting glucose (feature of insulin resistance), precedes type 2 diabetes mellitus and cardiovascular disease (5,6). In contrast, the hip circumference is associated with a decreased incidence in cardiovascular disease (7,8), diabetes mellitus (9,10), and might even predict health and longevity, especially in women (4).

Because diabetes mellitus is preceded by insulin resistance and lipid disturbances, we evaluated associations between body fat distribution (assessed by waist and hip circumference) and components of glucose metabolism, including clamp measurements of insulin sensitivity, triglyceride concentrations, and HDL-cholesterol.

Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

We analyzed data from the European Group of Insulin Resistance (EGIR) database (which comprises nondiabetic Caucasian Europeans) and the San Antonio Metabolism Study cohort (which includes non-Hispanic white Caucasian, and Hispanic residents of San Antonio, TX), including measurements of insulin sensitivity obtained with the use of the euglycemic insulin clamp technique (with an infusion rate of 240 pmol/min/m for 2 h). Details on the study cohort, the protocol, experimental and analytical methods have been published previously (11,12,13,14,15,16). Approval of the protocol by the local Ethics Committee and informed consent from all subjects was obtained before the studies at each geographical centre. For the purpose of the present study, insulin sensitivity was taken to be the steady-state total body insulin-mediated glucose disposal rate (MLBM), expressed as µmol per min per kg of fat-free (or lean body) mass (µmol/min/kg/lbm). Blood samples were collected in the fasting state for the measurement of fasting plasma glucose, fasting plasma insulin, fasting plasma triglyceride, and HDL-cholesterol. Plasma glucose was measured by the glucose oxidase method. Plasma insulin concentrations were measured by radioimmunoassay. Serum lipid levels were assayed by standard enzymatic assays. Fat mass (hence lean body mass) was measured by different techniques: underwater weighing, electrical bio-impedance, and tracer water. The results were generally in agreement, but for homogeneity it was decided to use the sex-specific Hume's formula throughout.

For the purpose of the present study, subjects without diabetes or impaired glucose tolerance were selected on basis of the availability of the anthropometric parameters BMI, waist, hip circumference, and insulin sensitivity; the steady-state total body insulin-mediated glucose disposal rate (M), expressed as µmol per min per kg of fat-free mass (µmol/min/kg lean body mass). Only subjects with complete data were analyzed (n = 439). Noneligible subjects of the database were in general younger and had a slightly lower BMI.

Statistical analyses

Data are given as means (± s.d.) or medians and their range, if distributions were skewed. Distributions of BMI, hip and waist circumference were divided into tertiles to provide insight into their mutual relationships and their relationship with insulin sensitivity and triglycerides, with the help of 3D graphs. Cut-points were calculated for male and female subjects separately, and the tertile-groups combined afterwards, to improve statistical power. Cut-points were 101 and 112 cm for hip circumference in female subjects (tertile-groups; lowest through 101, 101 through 112, 112 through highest), 99 and 104 cm for hip circumference in male subjects, 86 and 96 cm for waist circumference in female subjects, 90 and 98 cm for waist circumference in male subjects, 26 and 31 kg/m2 for BMI in female subjects, and 25 and 28 kg/m2 for BMI in male subjects. Differences between groups were analyzed by means of the Mann–Whitney U-test.

Linear regression analysis was applied to adjust for mutual confounding (hip and waist circumference), BMI, and age. Results were expressed as standardized coefficients (β) and R2, the proportion of variation “explained” by the independent variables. If appropriate, distributions were transformed into their natural logarithm to fulfill assumption criteria. A P value <0.05 was regarded statistically significant. Analyses were performed with the SPSS-PC software package, version 11.0.1 (SPSS, Chicago, IL).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

Table 1 shows population characteristics of 157 women and 282 men, selected from the EGIR and San Antonio study cohorts. The subjects had a wide range in age, BMI, waist and hip circumference. Data were provided in means (± s.d.) or median values and ranges, if distributions were skewed. Except for 44 Hispanics, subjects were Caucasians. Performing the analyses without the Hispanic population did not essentially change the results. Two women and 13 men had triglyceride concentrations >5.6 mmol/l and were excluded from analyses concerning triglycerides, though inclusion did not essentially change the results (data not shown). Triglyceride levels >5.6 mmol/l are clinically regarded as very high, and might be caused by specific lipid disorders (6). Table 1 showed further a wide range in BMI; performing analyses without subjects with a BMI >40 kg/m2 did also not essentially change the results (data not shown).

Table 1.  Study population characteristics, anthropometric measurements and biochemistry
inline image

Figure 1 shows that waist and hip circumference were strongly associated with BMI. BMI is lowest in subjects with a small hip and small waist circumference (1st tertile of both; 23 kg/m2) as compared to small waist (1st tertile) and large hip circumference (3rd tertile; 29 kg/m2, P < 0.001), or small hip (1st tertile) and large waist circumference (3rd tertile; 27 kg/m2, P < 0.001), etc. In the following analyses, all tertiles of BMI, waist and hip circumference were calculated separately for male and female subjects (see Methods and Procedures), and afterwards combined to improve statistical power.

image

Figure 1. Associations between BMI, hip and waist circumference in healthy subjects. BMI is lowest in subjects with a small hip and waist circumference (1st tertile of both; 23 kg/m2) as compared to small waist (1st tertile) and large hip circumference (3rd tertile; 29 kg/m2, P < 0.001) or small hip (1st tertile) and large waist circumference (3rd tertile; 27 kg/m2, P < 0.001). (BMI, hip and waist tertiles were calculated for male and female subjects separately).

Download figure to PowerPoint

In healthy subjects with a large BMI (3rd tertile), fasting triglyceride concentrations appeared to be lower in the presence of a large hip as compared to a small hip circumference (3rd vs. 1st tertile; 1.5 vs. 2.4 mmol/l, P < 0.038; Figure 2a). And, in subjects with similar BMI (3rd tertile), fasting triglyceride concentrations were higher in the presence of a large waist circumference vs. a small waist circumference (3rd vs. 1st tertile; 1.5 vs. 1.2 mmol/l, P < 0.025, Figure 2b).

image

Figure 2. Body fat distribution and triglycerides. (a) Association between triglycerides and hip circumference in healthy subjects. Triglyceride concentrations are lower in subjects with a large hip circumference (3rd tertile; 1.5 mmol/l) as compared to subjects with a small hip circumference (1st tertile; 2.4 mmol/l, P < 0.038), after stratification for BMI (hip and BMI tertiles were calculated for male and female subjects separately). (b) Association between triglycerides and waist circumference in healthy subjects. Triglyceride concentrations are higher in subjects with a large waist circumference (3rd tertile; 1.5 mmol/l) as compared to subjects with a small waist circumference (1st tertile; 1.2 mmol/l, P < 0.025), after stratification for BMI. (Waist and BMI tertiles were calculated for male and female subjects separately).

Download figure to PowerPoint

A similar analysis with respect to insulin sensitivity (clamp method), fasting glucose, fasting insulin, or HDL-cholesterol could not reveal such divergent associations of body fat distribution (data not shown, except for insulin sensitivity), i.e., in healthy subjects having a large BMI (3rd tertile), insulin sensitivity was not statistically different comparing a small hip circumference vs. a large hip circumference (1st vs. 3rd tertile; 42 vs. 33 µmol/min/kg lbm, P < 0.373; Figure 3a), also insulin sensitivity was not statistically different comparing a small waist circumference vs. a large waist circumference (1st vs. 3rd tertile; 40 vs. 33 µmol/min/kg lbm, P < 0.099; Figure 3b).

image

Figure 3. Body fat distribution and insulin sensitivity (clamp). (a) Association between insulin sensitivity and hip circumference in healthy subjects. A statistically significant difference in insulin sensitivity (clamp) could not be established between subjects with a small hip circumference (1st tertile; 42 µmol/min/kg lbm) as compared to subjects with a large hip circumference (3rd tertile; 33 µmol/min/kg lbm, P < 0.373), in the presence of a large BMI (3rd tertile). (Hip and BMI tertiles were calculated for male and female subjects separately). (b) Association between insulin sensitivity and waist circumference in healthy subjects. A statistically significant difference in insulin sensitivity (clamp) could not be established between subjects with a small waist circumference (1st tertile; 40 µmol/min/kg lbm) as compared to subjects with a large waist circumference (3rd tertile; 33 µmol/min/kg lbm, P < 0.099), in the presence of a large BMI (3rd tertile) (Waist and BMI tertiles were calculated for male and female subjects separately).

Download figure to PowerPoint

The divergent association of fasting triglyceride concentrations and body fat distribution was confirmed by use of linear regression analyses, allowing adjustment for BMI, mutual confounding (hip and waist circumference) and age. Analyses were performed for women and men separately, because body fat distribution generally shows distinct gender differences. Women showed an inverse association between fasting triglyceride concentrations and hip circumference (β ± s.e.: −0.78 ± 0.007, P < 0.001), and a (borderline statistically significant) positive association between fasting triglycerides and waist circumference (β ± s.e.: 0.24 ± 0.005, P < 0.065), after adjustment for BMI, mutual confounding, and age (Table 2). Similarly, men showed an inverse association between fasting triglyceride concentrations and hip circumference (β ± s.e.: −0.48 ± 0.005, P < 0.001), and a positive association between fasting triglycerides and waist circumference (β ± s.e.: 0.21 ± 0.005, P < 0.05), after adjustment for BMI, mutual confounding, and age.

Table 2.  Independent contributions of waist and hip circumference to components of lipid and glucose metabolism
inline image

Linear regression analyses did also reveal a divergent association between HDL-cholesterol and body fat distribution. Women showed a positive association between HDL-cholesterol and hip circumference (β ± s.e.: 0.39 ± 0.005, P < 0.05), and an inverse association between HDL-cholesterol and waist circumference (β ± s.e.: −0.35 ± 0.004, P < 0.01), after adjustment for mutual confounding, age, and BMI. Similarly, men showed a positive association between HDL-cholesterol and hip circumference (β ± s.e.: 0.36 ± 0.004, P < 0.001), and an inverse association between HDL-cholesterol and waist circumference (β ± s.e.: −0.32 ± 0.004, P < 0.001), after adjustment for mutual confounding, age, and BMI (Table 2).

Linear regression analyses, separated for women and men, did not reveal an association between insulin sensitivity and body fat distribution, though positive associations were shown in women for fasting glucose concentrations and waist circumference (β ± s.e.: 0.51 ± 0.005, P < 0.001), and positive associations were also shown for fasting insulin and waist circumference in women (β ± s.e.: 0.28 ± 0.005, P < 0.05), and men (β ± s.e.: 0.21 ± 0.004, P < 0.05) (Table 2). Replacement of BMI-linear by BMI-categorized showed essentially similar opposite associations for triglycerides and HDL-cholesterol, whereas it did not reveal opposite associations for glucose, insulin, or insulin sensitivity with body fat distribution. Furthermore, subanalyses of the European Group of Insulin Resistance (EGIR) database vs. the San Antonio Metabolism Study cohort did not essentially change the results, apart from loss of statistical power.

Repeating linear regression analyses, with combined data of men and women, did reveal a positive association between insulin sensitivity and hip circumference, though no statistically significant (P < 0.05) association with waist circumference, after adjustment for BMI, mutual confounding (hip and waist circumference), age, and gender. Further exploration, by evaluating effect modification in this model, revealed a statistically significant interaction between BMI and age on insulin sensitivity, which, however, did not essentially change the associations between hip and waist circumference and insulin sensitivity (data not shown). Yet, the variation (R2) of insulin sensitivity explained by hip circumference in these models was at maximum 1%. Replacement of insulin sensitivity by triglycerides in this model including the combined data of men and women, confirmed the findings of the separate analyses as provided in Table 2, and showed a highly significant inverse association between triglycerides and hip circumference, as well as a highly significant positive association between triglycerides and waist circumference. The variation (R2) of triglycerides explained by hip circumference was 9.5% and by waist circumference 6.5% (data not shown). Evaluation of effect modification between BMI and age in the latter model was not statistically significant. Evaluation of effect modification between BMI and hip or waist circumference, which might be suspected by the analyses depicted in Figure 2, was also not statistically significant (data not shown).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

We report here on a link between fasting triglyceride concentrations and body fat distribution in healthy persons. In persons generally regarded as having increased metabolic risk; the ones with a large BMI, triglyceride concentrations were lower in the presence of a large thigh (hip circumference), and higher in the presence of a large waist. The study is one of the first studies in humans to suggest, without proving, that adipocyte dysfunction precedes deterioration of muscle insulin sensitivity in obesity-related metabolic disturbances. The finding suggests a difference in triglyceride storage capacity between thigh and abdominal fat. This capacity to adequately store triglycerides may modulate the beneficial effects of gynoid obesity, and seems to play a more principal role than insulin sensitivity, in view of the fact that we did not find a link with insulin sensitivity in these healthy subjects, despite the use of the “gold standard method” to assess insulin sensitivity (17).

Thigh fat may contain adipose tissue better capable of buffering lipid-excess than abdominal fat (18). Recently, it was shown that women with a large thigh more efficiently stored their meal fat than women with a large amount of visceral fat (19). Maassen et al. anticipated that the buffering function of adipose tissue may depend on number and activity of mitochondria within adipocytes, contributing to the threshold at which fatty acids are released into the circulation, and, if deficient, leading to insulin resistance and type 2 diabetes mellitus (20,21). Preferential partitioning of lipid-excess into gluteal–femoral fat may directly affect glucose metabolism by reducing harmful effects of circulating free fatty acids on the ability of insulin to suppress hepatic glucose production (22), and it may also protect endothelial function (23).

The present analyses also revealed opposite associations between HDL-cholesterol and body fat distribution, after adjustment for confounding. Again, a large thigh was associated with favorable HDL-cholesterol concentrations whereas a large waist was associated with low cholesterol concentrations. HDL-cholesterol appears to play an important role with respect to the development of type 2 diabetes mellitus as well as cardiovascular disease (24,25). Triglycerides and HDL-cholesterol are inversely associated, though triglycerides may more clearly explain their contribution to the development of metabolic disturbances (26).

The findings on triglyceride- and HDL levels are somehow in agreement with former reports, though the absence of a statistically significant opposite association between insulin sensitivity and body fat distribution in the same persons with similar analytic techniques makes the present study distinctive (27,28,29,30). Of course, if more persons had been included in the study opposite associations with insulin sensitivity could have been revealed, though the finding that opposite associations with triglycerides are far stronger and already detectable in simple stratified analyses (Figure 2), as well as the fact that 16% of variation in triglycerides (R2) was explained by hip and waist circumference vs. at maximum 1% of variation of insulin sensitivity (R2) by hip circumference alone (in linear regression models), supports a primary role for triglycerides in the metabolic risk difference between hip and waist circumference, and is in line with a recent prospective study in healthy young men, reporting prediction of diabetes by an increase in triglyceride level over time (31). Other studies on the topic showed variable findings. One study showed favorable associations between leg fat and parameters of insulin resistance, as calculated by values of glucose and insulin after an oral glucose tolerance test (28). Another study showed a positive association between lower-body fat mass and insulin sensitivity as estimated by the Matsudas index; they concluded that this effect could partly be explained by variations in serum adiponectin levels (29). They reported also lower serum insulin in a subgroup of men selected for large lower-body fat mass concentration as compared to a subgroup of men selected with small lower-body fat mass, matched for trunkal fat mass. Given that we used most precise measurements of insulin sensitivity in a healthy population (17), the better associations with triglycerides and HDL-cholesterol as compared to insulin sensitivity suggest an initial disturbance of lipid metabolism, which partly depend on body fat distribution, though longitudinal studies are required to find out whether disturbances in lipid metabolism precede insulin resistance. Thus, part of the discrepancy might be explained by different techniques to assess insulin sensitivity and by different study-populations, with consequently different stages in the pathogenic pathway. One study was performed in healthy postmenopausal women (28), and another in randomly selected men in combination with juvenile onset obese men (29). Two large studies showed opposite effects on triglycerides and HDL-cholesterol of waist vs. hip circumference in agreement with our results (27,30), though both studies showed also consistently opposite effects on glucose and insulin. One of these studies included patients with diabetes (27), and the other study included subjects with impaired glucose tolerance (30). Consequently, the range of glucose and insulin is much larger in these populations than in our healthy population (without diabetes and impaired glucose tolerance), which facilitates detection of an association. The absence of an opposite association of insulin sensitivity with body fat distribution in the present study might therefore be even more intriguing, because the range in insulin sensitivity is large in the present healthy study population (11).

Definite conclusions cannot be made with respect to causal interference, because most of the available data are cross-sectional (27,28,29,30), and measurements of one prospective study are limited (31). It is further not completely clear whether the association between the lipids and body fat distribution is linear or more explicit in subjects with a large BMI; further studies in large healthy populations are needed. Another potential limitation of the present study is that hip circumference was not separated into peripheral muscle, fat masses, and pelvic size, because peripheral muscle atrophy might influence insulin sensitivity (32). Small differences between women and men have not been evaluated; they might be explained by regional differences in adipocyte metabolism with respect to lipoprotein lipase activity and lipolysis, or different effects of adrenal hormones and sex steroids (33).

In conclusion, hip and waist circumference show divergent effects with respect to fasting triglycerides and HDL-cholesterol. The findings support a protective role by gynoid adiposity through better triglyceride storage capacity, preventing harmful lipid exposure.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES

We are thankful for the generous permission of E. Ferrannini and R.A. Defronzo to conduct the present study with data of the European Group for the Study of Insulin Resistance (EGIR) and San Antonio.

REFERENCES

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Disclosure
  9. REFERENCES
  • 1
    Haffner SM, Taegtmeyer H. Epidemic obesity and the metabolic syndrome. Circulation 2003; 108: 15411545.
  • 2
    Lean MJ, Han TS. Waist worries. Am J Clin Nutr 2002; 76: 699700.
  • 3
    Wahrenberg H, Hertel K, Leijonhufvud BM et al. Use of waist circumference to predict insulin resistance: retrospective study. BMJ 2005; 330: 13631364.
  • 4
    Lissner L, Bjorkelund C, Heitmann BL, Seidell JC, Bengtsson C. Larger hip circumference independently predicts health and longevity in a Swedish female cohort. Obes Res 2001; 9: 644646.
  • 5
    World Health Organization. Definition, Diagnosis and Classification of Diabetes Mellitus and its Complications: Report of a WHO Consultation. Geneva, 1998. WHO/NCD/NCS 99.2.
  • 6
    National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation 2002; 106: 31433421.
  • 7
    Tanko LB, Bagger YZ, Alexandersen P, Larsen PJ, Christiansen C. Peripheral adiposity exhibits an independent dominant antiatherogenic effect in elderly women. Circulation 2003; 107: 16261631.
  • 8
    Heitmann BL, Frederiksen P, Lissner L. Hip circumference and cardiovascular morbidity and mortality in men and women. Obes Res 2004; 12: 482487.
  • 9
    Seidell JC, Han TS, Feskens EJ, Lean ME. Narrow hips and broad waist circumferences independently contribute to increased risk of non-insulin-dependent diabetes mellitus. J Intern Med 1997; 242: 401406.
  • 10
    Snijder MB, Dekker JM, Visser M et al. Associations of hip and thigh circumferences independent of waist circumference with the incidence of type 2 diabetes: the Hoorn Study. Am J Clin Nutr 2003; 77: 11921197.
  • 11
    Ruige JB, Mertens IL, Bartholomeeusen E et al. Fasting-based estimates of insulin sensitivity in overweight and obesity: a critical appraisal. Obesity 2006; 14: 12501256.
  • 12
    Natali A, Toschi E, Camastra S et al. Determinants of postabsorptive endogenous glucose output in non-diabetic subjects. European Group for the Study of Insulin Resistance (EGIR). Diabetologia 2000; 43: 12661272.
  • 13
    Ferrannini E, Natali A, Capaldo B et al. Insulin resistance, hyperinsulinemia, and blood pressure: role of age and obesity. European Group for the Study of Insulin Resistance (EGIR). Hypertension 1997; 30: 11441149.
  • 14
    Ferrannini E, Natali A, Bell P et al. Insulin resistance and hypersecretion in obesity. European Group for the Study of Insulin Resistance (EGIR). J Clin Invest 1997; 100: 11661173.
  • 15
    Ferrannini E, Vichi S, Beck-Nielsen H et al. Insulin action and age. European Group for the Study of Insulin Resistance (EGIR). Diabetes 1996; 45: 947953.
  • 16
    Baldeweg SE, Golay A, Natali A et al. Insulin resistance, lipid and fatty acid concentrations in 867 healthy Europeans. European Group for the Study of Insulin Resistance (EGIR). Eur J Clin Invest 2000; 30: 4552.
  • 17
    Ferrannini E, Mari A. How to measure insulin sensitivity. J Hypertens 1998; 16: 895906.
  • 18
    Frayn KN. Adipose tissue as a buffer for daily lipid flux. Diabetologia 2002; 45: 12011210.
  • 19
    Votruba SB, Mattison RS, Dumesic DA, Koutsari C, Jensen MD. Meal fatty acid uptake in visceral fat in women. Diabetes 2007; 56: 25892597.
  • 20
    Maassen JA, Romijn JA, Heine RJ. Fatty acid-induced mitochondrial uncoupling in adipocytes as a key protective factor against insulin resistance and beta cell dysfunction: a new concept in the pathogenesis of obesity-associated type 2 diabetes mellitus. Diabetologia 2007; 50: 20362041.
  • 21
    Maassen JA, Hart LM 't, Ouwens DM. Lessons that can be learned from patients with diabetogenic mutations in mitochondrial DNA: implications for common type 2 diabetes. Curr Opin Clin Nutr Metab Care 2007; 10: 693697.
  • 22
    Boden G. Role of fatty acids in the pathogenesis of insulin resistance and NIDDM. Diabetes 1997; 46: 310.
  • 23
    Ferreira AC, Peter AA, Mendez AJ et al. Postprandial hypertriglyceridemia increases circulating levels of endothelial cell microparticles. Circulation 2004; 110: 35993603.
  • 24
    Fagot-Campagna A, Knowler WC, Narayan KM et al. HDL cholesterol subfractions and risk of developing type 2 diabetes among Pima Indians. Diabetes Care 1999; 22: 271274.
  • 25
    Rizzo M, Corrado E, Coppola G et al. Prediction of cardio- and cerebro-vascular events in patients with subclinical carotid atherosclerosis and low HDL-cholesterol. Atherosclerosis 2008; 200: 389395.
  • 26
    Lewis GF, Carpentier A, Adeli K, Giacca A. Disordered fat storage and mobilization in the pathogenesis of insulin resistance and type 2 diabetes. Endocr Rev 2002; 23: 201229.
  • 27
    Seidell JC, Perusse L, Despres 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: 315321.
  • 28
    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: E1023E1028.
  • 29
    Buemann B, Sorensen TI, Pedersen O et al. Lower-body fat mass as an independent marker of insulin sensitivity—the role of adiponectin. Int J Obes Relat Metab Disord 2005; 29: 624631.
  • 30
    Snijder MB, Zimmet PZ, Visser M et al. Independent association of hip circumference with metabolic profile in different ethnic groups. Obes Res 2004; 12: 13701374.
  • 31
    Tirosh A, Shai I, Bitzur R et al. Changes in triglyceride levels over tme and risk of type 2 diabetes in young men. Diabetes Care 2008; 31: 20322037.
  • 32
    Koh-Banerjee P, Wang Y, Hu FB et al. Changes in body weight and body fat distribution as risk factors for clinical diabetes in US men. Am J Epidemiol 2004; 159: 11501159.
  • 33
    Fried SK, Russell CD, Grauso NL, Brolin RE. Lipoprotein lipase regulation by insulin and glucocorticoid in subcutaneous and omental adipose tissues of obese women and men. J Clin Invest 1993; 92: 21912198.