Is visceral obesity the cause of the metabolic syndrome?

Abstract

Despite the fact that controversy remains around the underlying pathophysiological processes leading to the development of the metabolic syndrome (insulin resistance and/or hyperinsulinemia versus abdominal obesity), there is increased recognition that abdominal obesity is the most prevalent form of the metabolic syndrome. Although it has been well established that there is a greater prevalence of chronic metabolic diseases such as diabetes and cardiovascular diseases in obese patients than among normal weight individuals, obesity is a remarkably heterogeneous condition and not every obese patient is characterized by co‐morbidities. In this regard, body fat distribution, especially visceral adipose tissue accumulation, has been found to be a major correlate of a cluster of diabetogenic, atherogenic, prothrombotic and proinflammatory metabolic abnormalities referred to as the metabolic syndrome. Due to its anatomic location and peculiar metabolic, hyperlipolytic activity, the expanded visceral adipose depot is a key correlate of the altered cardiometabolic risk profile observed among individuals with a high‐risk abdominal obesity phenotype. Evidence suggests that this dysmetabolic profile is predictive of a substantially increased risk of coronary heart disease even in the absence of classical risk factors. Finally, a moderate weight loss in initially abdominally obese patients is associated with a preferential mobilization of visceral adipose tissue, which in turn leads to substantial improvements in the metabolic risk profile predictive of a reduced risk of coronary heart disease and of type 2 diabetes.

• Adipose tissue
• cardiometabolic risk profile
• clinical screening tools
• metabolic syndrome
• visceral obesity

Introduction

In 1988, Reaven 1 coined the term syndrome X and proposed that an impaired in vivo insulin action was a core component responsible for a cluster of abnormalities/risk factors which included compensatory hyperinsulinemia, hypertriglyceridemia, low high‐density lipoprotein (HDL)‐cholesterol levels and hypertension (Figure 1A). As the simultaneous presence of these metabolic abnormalities has been shown to increase the risk of coronary heart disease (CHD) even in the absence of diabetes or hyperglycemia 2, the introduction of such a concept was a remarkable advance in preventive medicine as it did emphasize the need to go beyond the assessment of conventional risk factors in order to identify high‐risk individuals. At the same time that Reaven's syndrome X was introduced, DeFronzo 3 underlined the importance of a triad of skeletal muscle, liver and beta‐cell abnormalities in the etiology of type 2 diabetes. He did emphasize that skeletal muscle insulin resistance was an essential element in the pathophysiology of type 2 diabetes. Under his model, an insulin resistant liver leading to an increased hepatic glucose production would contribute to explain the glucose intolerance of insulin resistant individuals. As insulin resistance can be frequently observed in non‐diabetic individuals, DeFronzo also emphasized that a relative deficit in beta‐cell insulin secretion capacity was essential to observe the conversion from an insulin resistant state to type 2 diabetes. Since it has now become obvious that an insulin resistant state is also accompanied by an impaired antilipolytic effect of insulin in adipose tissue leading to an overproduction of free fatty acids (FFA) 4,5, DeFronzo's ‘triumvirate’ concept has evolved to the notion of a ‘harmonious quartet’ 6 where skeletal muscle insulin resistance, increased hepatic glucose production, increased adipose tissue lipolysis and reduced beta‐cell insulin secretion capacity are all involved in the regulation of the hyperglycemic state of type 2 diabetes (Figure 1B).

Figure 1. A. Insulin resistance syndrome or syndrome X as initially described by Reaven in 1988. B. DeFronzo's ‘harmonious quartet’ showing abnormalities involved in the pathophysiology of type 2 diabetes.

As insulin resistance can be measured in specialized clinical research settings even in the absence of hyperglycemia, and as there is evidence that markers of an insulin resistant state are predictive of an increased risk of CHD 7, there has been considerable interest in the study of the clustering abnormalities observed among insulin resistant individuals. Furthermore, the pathophysiology of insulin resistance has been the topic of excellent review papers 8,9 and the study of the possible molecular mechanisms involved in the etiology of insulin resistance has been and remains the subject of numerous investigations. Addressing this specific topic is beyond the scope of the present paper and the reader is referred to this previously published literature. As there is increased recognition that abdominal obesity is by far the most prevalent form of insulin resistance 10, the present paper will focus on the possible mechanisms linking abdominal obesity to insulin resistance and to its related cluster of atherogenic metabolic abnormalities which are now often referred to as the metabolic syndrome (Figure 2).

Figure 2. Distinction to be made between the conceptual definition of the most prevalent form of the metabolic syndrome(abdominal obesity) (panel A) versus clinical tools that have been proposed by various organizations/groups to identify patients likely to have the metabolic syndrome (panel B). Although finding patients with the metabolic syndrome is important in clinical practice, how to integrate this information in global cardiovascular risk assessment remains an unresolved issue (see text). NCEP‐ATP III = National Cholesterol Education Program‐Adult Treatment Panel III; IDF = International Diabetes Federation; WHO = World Health Organization; EGIR = European Group for the study of Insulin Resistance; AACE = American Association of Clinical Endocrinologists; HyperTG waist = hypertriglyceridemic waist; BMI = body mass index; WHR = waist‐to‐hip ratio. ^*Insulin resistance measured under hyperinsulinemic euglycemic conditions. ^†Includes family history of type 2 diabetes, polycystic ovary syndrome, sedentary lifestyle, advancing age and ethnic groups susceptible to type 2 diabetes.

Key messages

• Abdominal obesity is the most prevalent form of the metabolic syndrome.

• Adipose tissue is also an endocrine organ which secretes adipokines involved in the atherogenic/diabetogenic metabolic risk profile of abdominal obesity.

• Simple and discriminant tools must be developed to optimally assess the global cardiovascular risk associated with the presence of classical risk factors and markers of the metabolic syndrome (cardiometabolic risk).

Obesity, insulin resistance and the metabolic syndrome: Beyond body weight

Despite the well documented health hazards of obesity 11,12, physicians are constantly challenged by the absence of complications in some very obese patients who may nevertheless have a fairly normal metabolic risk factor profile despite their very significant obesity. On the other hand, some moderately overweight individuals are characterized by a whole cluster of atherogenic and diabetogenic metabolic abnormalities. In this regard, a whole series of imaging and metabolic studies conducted over almost two decades have revealed that, in both men and women, the amount of visceral adipose tissue, which can be reliably assessed by imaging techniques such as computed tomography, is a critical correlate of the presence of a substantially altered metabolic risk factor profile predictive of an ncreased risk for the development of type 2 diabetes and cardiovascular disease (CVD) 13–18. Therefore, although there is evidence that obesity increases the risk of chronic complications, there is also compelling evidence that viscerally obese patients represent the subgroup of obese subjects characterized by the most severe metabolic abnormalities.

Visceral obesity, insulin resistance and additional features of the metabolic syndrome: The portal theory and beyond

There is ample evidence indicating that an increased adipose tissue mass is associated with an impaired in vivo insulin action. Such association had led some investigators to suggest that subcutaneous and not visceral adiposity was the key correlate of insulin resistance 19–22. Although there is solid evidence that the small visceral adipose depot does not make a major contribution to systemic FFA levels 23, studies that have carefully matched subjects for total body fat (or for subcutaneous fat) but with high versus low levels of visceral adipose tissue have clearly shown that viscerally obese patients have the most disturbed metabolic risk profile 14,24–26. Furthermore, the notion that subcutaneous fat is the culprit cannot be reconciled with the observation in clinical practice of very obese patients who have a normal metabolic risk profile when they have low levels of visceral adipose tissue 27–29. Thus, studies that have used imaging techniques to measure subcutaneous versus visceral adipose tissue accumulation have clearly shown that viscerally obese subjects (even in the absence of clinical obesity) represented the subgroup of individuals characterized by the most severe insulin resistant state 14,24,25,30. Because of the peculiar metabolism and anatomic location of the intra‐abdominal adipose tissue which is partly drained by the portal vein, an increased accumulation of hyperlipolytic visceral adipose tissue has therefore been associated with an impaired FFA metabolism (as these hypertrophied visceral adipocytes have been shown to be resistant to the antilipolytic effect of insulin and also characterized by an impaired FFA esterification) 31,32 (Figure 3). Such flux of FFA toward the liver favors an increased secretion of triglyceride (TG)‐rich lipoproteins, reduced hepatic degradation of apolipoprotein B and of insulin leading to hyperapolipoprotein B and hyperinsulinemia, and to an increased hepatic glucose production contributing to glucose intolerance 33,34. In addition, it has been reported that viscerally obese patients are characterized by an exaggerated postprandial TG response 17 as the increased very low‐density lipoprotein (VLDL) levels associated with a fasting hypertriglyceridemic state compete with dietary chylomicrons for their lipolysis by the enzyme lipoprotein lipase which is saturated in the postprandial state. Such an elevated concentration of TG‐rich lipoproteins found in viscerally obese patients favors the transfer of TG from TG‐rich lipoproteins to low‐density lipoprotein (LDL) and HDL in exchange for cholesteryl esters located in LDL and HDL. This process leads to the TG‐enrichment of LDL and HDL particles which are then subjected to lipolysis by the enzyme hepatic lipase. There is some evidence suggesting that hepatic lipase activity is elevated in viscerally obese patients 35,36 and an increased activity of this enzyme promotes the formation of small LDL and HDL particles. As the core of LDL and HDL particles are relatively depleted in esterified cholesterol, the lipolysis of the core‐TG molecules thus leads to the formation of cholesteryl ester‐depleted, small LDL and HDL particles. Such process explains the increased proportion of small LDL and HDL particles in visceral obesity and is also largely responsible for the reduced HDL‐cholesterol concentration found in this condition 18,37.Thus, we have reported that the small, dense LDL and HDL phenotype is also a component of the high TG, reduced HDL‐cholesterol dyslipidemia which is found in abdominally obese insulin resistant patients (Figure 4) 38. In this regard, plasma LDL‐cholesterol concentration is not increased in viscerally obese patients 15,26. In addition, as the plasma LDL‐cholesterol concentration shows no correlation with LDL and HDL particle sizes 13,38, clinicians should not rely on LDL‐cholesterol levels to identify patients likely to have small LDL or HDL particles. Thus, there is a metabolic rationale linking the expanded abdominal visceral depot to liver insulin resistance, hyperinsulinemia, glucose intolerance and to the high TG, low HDL‐cholesterol, small LDL dyslipidemia often observed in abdominally obese patients.

Figure 3. Working mechanistic model for the relation of excess visceral adipose tissue accumulation to the insulin resistant, dyslipidemic, prothrombotic and proinflammatory profile of abdominally obese patients.

Figure 4. Triglyceride and HDL‐cholesterol levels in men according to LDL particle size (smaller<251.6 Å; larger ⩾251.6 Å) and HDL particle size (smaller<83.5 Å; larger ⩾83.5 Å). 1,2,3: significantly different from the corresponding subgroup P<0.0001. Thus, subjects with both small LDL and small HDL particles had the highest plasma triglycerides and the lowest HDL‐cholesterol concentrations.

However, this ‘portal theory’ has often been questioned 39,40 and evidence has also been provided that such a model cannot reconcile all metabolic abnormalities found in abdominally obese patients. For example, even among viscerally obese patients, the expanded visceral adipose depot does not represent a large proportion of the total adipose tissue mass 39. Thus, even if visceral adipocytes have a hyperlipolytic state, a large percentage of adipose tissue‐derived fatty acids must be derived from the subcutaneous fat depots 41. Another possibility is that visceral obesity is a marker of the relative inability of subcutaneous fat to clear and accumulate the energy surplus originating from the diet 42. It has also been proposed that visceral obesity could be a marker of a more primary neuroendocrine abnormality preferentially partitioning the energy surplus in visceral rather than in subcutaneous adipose tissue 43,44. For instance, the higher glucocorticoid receptor density in visceral than subcutaneous fat cells could make the visceral depot exquisitely sensitive to an activated hypothalamic‐pituitary‐adrenal axis, leading to visceral adipose tissue accumulation 44,45. An obvious clinical example of this situation is Cushing's syndrome, in which patients have both high cortisol levels and expanded waistlines as well as evidence of marked insulin resistance 46. Steroid hormones are also obviously involved as estrogens increase peripheral fat accumulation in women whereas increased androgen levels appear to protect against visceral fat deposition in men 47. Metabolic and body fat distribution studies conducted on transsexuals who received hormone replacement therapy have clearly shown the remarkable impact of sex steroids on body fat distribution and related metabolic parameters 48,49. Thus, visceral obesity may be a marker of an abnormal neuro‐endocrine profile leading to both features of insulin resistance and to visceral obesity. However, once the visceral depot is expanded, the peculiar metabolic profile of the hypertrophied visceral adipocytes may exacerbate the patient's condition and the individual may enter a vicious cycle exacerbating his/her risk of diabetes and CHD (Figure 5).

Figure 5. Hypothetical model by which visceral obesity is partly the consequence of a given neuro‐hormonal milieu preferentially partitioning the excess energy in the visceral fat depot rather than in subcutaneous adipose tissue.

Adipose tissue as an endocrine organ: The missing link?

It is also now recognized that adipose tissue is a remarkable endocrine organ which produces pro‐thrombotic and inflammatory molecules 50. Such ‘adipokines’ could contribute to exacerbate the patient's risk of diabetes and CHD. For instance, viscerally obese patients have been shown to have an impaired fibrinolysis and an increased susceptibility to thrombosis as well as evidence for the presence of a chronic inflammatory state 16, 51–56. Markedly elevated plasma C‐reactive protein (CRP) concentrations have been reported in abdominally obese patients, the highest levels being found in obese patients with a selective excess of visceral adipose tissue 16. Such elevated CRP levels may be the consequence of altered cytokine levels (elevated interleukin (IL)‐6 and tumor necrosis factor (TNF)‐α originating from the expanded adipose tissue) 57,58 promoting a chronic inflammatory state which could also contribute to the etiology of insulin resistance and CVD. Recently, another adipose tissue‐specific cytokine, adiponectin, has been reported to be reduced in abdominally obese patients 59,60. Such a low adiponectin concentration has been suggested to contribute to the development of insulin resistance and CHD in abdominally obese patients with the features of the metabolic syndrome 59,61,61. Cross‐sectional studies have reported lower adiponectin levels in type 2 diabetic patients as well as in patients with documented CHD 63,64. Furthermore, a recent prospective study has shown that a low adiponectin concentration could predict an increased CHD risk beyond the contribution of classical CHD risk factors 65. Several mechanisms could explain the apparently favorable effects of adiponectin regarding its potential protection against type 2 diabetes and CHD and they are summarized in Table I.

Table I. In vitro metabolic effects of adiponectin which may potentially protect against the development of type 2 diabetes and coronary heart disease.

Anti‐atherogenic properties:

↓ expression of adhesion molecules

↓ monocyte adhesion to endothelial cells

↓ uptake of oxidized LDL

↓ foam cell formation

↓ proliferation and migration of smooth muscle cells

Anti‐diabetic properties:

↑ insulin sensitivity

↑ smooth muscle glucose uptake and free fatty acid oxidation

↓ hepatic glucose production

↓ intracellular triglycerides

Therefore, viscerally obese patients, even in the absence of classical CVD risk factors such as hypertension, hyperglycemia, elevated LDL‐cholesterol concentrations, could be at increased risk for an acute coronary syndrome due to the presence of an atherogenic, prothrombotic and proinflammatory profile 38. However, although several features of this metabolic syndrome have been shown to predict an increased CHD risk, we presently do not know which features of the metabolic syndrome are critical to assess in order to improve our ability to predict CHD events once traditional risk factors are taken into account. This issue is the topic of considerable debate in the literature 66. Additional studies are clearly warranted and this will be a fertile area of investigation in future studies.

The cluster of metabolic abnormalities of visceral obesity: Impact on CHD risk

Over the years, attempts have been made to quantify the risk of CHD associated with the features of the metabolic syndrome that we and others have found in abdominally obese men, even in the absence of type 2 diabetes. Several prospective studies have suggested that fasting hyperinsulinemia, as a crude marker of insulin resistance in non‐diabetic individuals, was predictive of an increased CHD risk 67–71. In the sample of 2103 asymptomatic middle‐aged men of the Québec Cardiovascular Study followed for incidence of a first CHD event over 5 years, we found that some of the metabolic features of abdominal obesity, namely fasting hyperinsulinemia (as a crude marker of insulin resistance), elevated apolipoprotein B concentration, small LDL particles, an elevated cholesterol/HDL‐cholesterol ratio, and an elevated CRP concentration (a marker of low chronic inflammation) were all predictive of an increased risk of CHD 68,72–75. We also found that the simultaneous presence of these features of the metabolic syndrome were predictive of a markedly elevated risk of CHD 7. As some of these markers are not well accepted CHD risk factors in current medical practice and in national guidelines, we have compared the ability of some of these metabolic markers of the insulin resistance syndrome (the atherogenic metabolic triad of elevated insulin and apolipoprotein B levels and small LDL particles) to predict CHD risk as opposed to commonly used algorithms such as the Framingham risk chart 76. Not surprisingly, we found that a high Framingham score was predictive of an increased risk of CHD in men of the Québec Cardiovascular Study 76. However, even in the absence of classical risk factors (and therefore among men with a low Framingham score), we found the features of the metabolic syndrome to be predictive also of an equally increased risk of CHD 76. Therefore, these results indicate that our current risk assessment algorithm based on Framingham risk factors does not always identify high‐risk individuals with the features of the metabolic syndrome if these individuals do not have traditional risk factors such as hypertension, smoking, elevated cholesterol or type 2 diabetes. There is therefore a need to further refine current CHD risk assessment algorithms by incorporating critical markers of the metabolic syndrome. However, which combination of markers would provide the optimal discrimination of high‐risk individuals at the lowest cost while being easily standardized remains an unresolved issue.

This is essentially the rationale which was used to develop the ‘hypertriglyceridemic waist’ concept 77 as there was a need to implement in clinical practice a simple screening approach which could help primary care physicians to identify, in a cost‐effective manner, the subgroup of abdominally obese patients with the atherogenic features of the metabolic syndrome. ‘Hypertriglyceridemic waist’ must also be discussed in the context of guidelines published to identify patients likely to have the metabolic syndrome. For instance, the National Cholesterol Education Program‐Adult Treatment Panel III (NCEP‐ATP III) guidelines 78 have also acknowledged the importance of considering waist girth and fasting TG as two key elements to identify individuals with the metabolic syndrome although different cutoff values were proposed. Additional work will be required to validate the screening tools proposed by various organizations. It is recommended that inclusion of variables and selection of cutoff values be based on hard clinical events. Lack of validation of clinical criteria of the metabolic syndrome against ‘hard’ CHD events is one of the criticisms addressed 66 to the metabolic syndrome criteria/cutoffs proposed by current guidelines 77–82 (Figure 2B). Furthermore, once global CHD risk is assessed by taking into account ‘classical’ risk factors, it has been claimed that whether the presence/absence of clinical criteria of the metabolic syndrome would alter clinical management of these patients is not clear 83. However, once the notion that the most prevalent form of the metabolic syndrome is abdominal obesity, such a concept emphasizes the need to assess and target abdominal obesity, an objective poorly achieved in current clinical practice.

Targeting weight loss in high‐risk abdominally obese patients with the metabolic syndrome: Recent evidence and unanswered questions

Despite the evidence that abdominal obesity characterized by an excess of visceral adipose tissue and related features of the metabolic syndrome is a high‐risk obesity phenotype, no study has shown that aiming at visceral adipose tissue loss as a therapeutic target would reduce the risk of CHD and of type 2 diabetes in patients characterized by this condition. The best data available on the long‐term benefits of moderate weight loss come from two landmark trials that were recently published: the Finnish Prevention of Diabetes study 84 and the American Diabetes Prevention Program 85. Both trials recruited non‐diabetic patients with impaired glucose tolerance (IGT) who were randomized either to a lifestyle modification program aiming at moderate weight loss (produced by a combination of increased physical activity and moderate reduction of caloric intake) or to a placebo group receiving usual medical care. In the American Diabetes Prevention Program, there was also another group who received metformin. Both trials were remarkably similar in showing that moderate weight loss induced by the lifestyle modification program reduced the incidence of diabetes by 58%. Furthermore, the American Diabetes Prevention Program showed that the risk of conversion from IGT to diabetes was reduced to a greater extent (58%) by lifestyle than by metformin (31%). Another trial that used a weight loss agent (orlistat) also reported the added benefits of pharmacotherapy of obesity beyond a lifestyle modification program in reducing the risk of developing type 2 diabetes over 4 years, such benefits being particularly important among obese patients with IGT baseline 86.

The above three trials provide evidence that moderate weight loss produced by an improved diet and a more physically active lifestyle (without/with pharmacotherapy) can indeed decrease the risk of developing type 2 diabetes in a high‐risk population with IGT. However, although the contribution of these studies is remarkable in showing that prevention can work and in providing evidence on the type of support and of facilities needed to successfully induce moderate weight loss in subjects at high risk for type 2 diabetes, several questions remain unanswered. Although no computed tomography data on visceral adiposity are available, subjects enrolled in the Finnish Prevention of Diabetes study and the American Diabetes Prevention Program were abdominally obese. Thus, was the substantial reduction in risk of developing diabetes related to a selective loss of visceral adipose tissue? Was the assumed loss of visceral adipose tissue maintained throughout the two trials? Would the loss of visceral adipose tissue induced by moderate weight loss be maintained once a reduced body weight is achieved and maintained over time? Did moderate weight loss prevent or just retard conversion to diabetes? Is such a moderate weight loss sufficient to substantially improve the features of the metabolic syndrome in patients with IGT? Would the assumed improvements in metabolic risk factors produced by these lifestyle modification programs be sustained had the reduced body weight been maintained over time?

Clearly, a moderate weight loss of only 5%–10% is associated with greater metabolic benefits than what could be explained by such fairly limited weight loss which does not normalize body weight in most obese patients 87,88. There is a possible explanation for this phenomenon. For instance, we and others have shown that weight loss is associated with a preferential mobilization of visceral adipose tissue compared to subcutaneous fat, especially among patients with a high accumulation of visceral adipose tissue at baseline 88–93. From the close link between visceral obesity and metabolic complications, such a loss of visceral fat may generate substantial improvements in the metabolic risk factor profile, a concept concordant with results of Goodpaster et al. 90 and of Ross et al. 94,95 who have shown that the improvements in insulin sensitivity/glucose tolerance in response to weight loss were more closely related to the loss of visceral adipose tissue than to the reduction in total body fat or subcutaneous fat.

In summary, there is considerable evidence that the most prevalent form of the metabolic syndrome is associated with abdominal obesity, especially when accompanied by a selective deposition of visceral adipose tissue. This notion has important clinical implications as it does recognize the importance of identifying the expanded waist girth of metabolic syndrome patients as a therapeutic target for the optimal management of their CHD risk. Although most physicians have heard about the metabolic syndrome and now recognize its importance, they are nevertheless confused on how to detect patients with the metabolic syndrome, how to optimally assess global CHD risk and how to manage these patients, with proper identification of key variables and targets. The current debate about the relevance of considering the metabolic syndrome in clinical practice 66,83 highlights the fact that physicians often confuse pathophysiology and clinical screening tools. While a considerable amount of additional research is needed to understand which are the key features of the metabolic syndrome that really add to global CHD risk as assessed by classical risk factors, simple and discriminant tools must be developed and provided to clinicians as there is an urgent need to implement efficient screening strategies to identify high‐risk abdominally obese patients.

Finally, despite our remarkable progress in our ability to treat CVD risk factors by pharmacotherapy, our clinical efficacy is quite limited and there is evidence that a substantial proportion of treated patients remain at high risk for an acute coronary event. It is proposed that until abdominal obesity is identified as a key therapeutic target, we will not optimally manage the risk of CVD in a clinically significant proportion of high‐risk patients 96.

Acknowledgments

Dr. Després' work has been supported by research grants from the Canadian Institutes of Health Research and the Canadian Diabetes Association.