Multiple adipose depots increase cardiovascular risk via local and systemic effects

Abstract

Adipose tissue modifies the development of cardiovascular disease in a complex manner: obesity is a major risk factor, but particularly when accompanied with a central fat distribution. For that reason the characteristics of visceral adipose tissue attracted the majority of research interest thus far and measurement of waist circumference is now recommended for everyday clinical practice. However, the direct, causative role of visceral fat in cardiometabolic disease remains to be established. Epidemiological and clinical studies show that accumulation of fat subcutaneously, in the gluteo-femoral area, is protective for cardiovascular disease, but the exact molecular mechanisms remain again unclear. In the last few years, imaging allowed the study of smaller fat depots that may interact locally with important tissues: epicardial fat with the myocardium, perivascular fat with the vessel wall and the developing atherosclerotic plaque, renal sinus fat with the renal artery. Unraveling the heterogeneous fat distribution and metabolic phenotypes in human obesity will facilitate optimal assessment of cardiovascular risk in overweight and obese individuals.

Introduction

The association between obesity and cardiovascular risk has been described in multiple epidemiological studies and is by now universally accepted [1]. However, this is far from a straightforward relationship; for any given body mass index (BMI) level, there is significant variability with some lean individuals developing disease and others remaining healthy despite severe obesity, the so-called metabolically healthy obese (MHO) [2]. A significant part of this variability can be attributed to variation in fat distribution and depot differences in adipose tissue function.

Adipose tissue is unique in that it is dispersed throughout the body in discreet depots, which are thought to constitute “separate mini-organs” [3]. Almost 90% of the adipose tissue is found subcutaneously (subcutaneous adipose tissue – SAT), mainly in the abdominal, gluteal and femoral areas. Visceral adipose tissue (VAT) and smaller depots in close proximity to the heart (epicardial adipose tissue – EAT), kidneys, joints, eyes, etc make up the remaining 10-20% [4]. The size of each depot and the balance between them is important for the cardiometabolic risk of the individual: a peripheral fat distribution (in the limbs) is favorable and a central one (truncal, both subcutaneous and visceral) detrimental as will be discussed in more detail below. The mechanisms linking fat distribution and cardiovascular risk are complex and involve changes in whole body glucose and lipid metabolism, effects on traditional and novel cardiovascular risk factors (e.g. hypertension, inflammation) and the systemic or local actions of adipokines secreted by the adipose tissue.

Epidemiological and imaging studies link adipose tissue depots and atherosclerosis risk, severity and progression

Central obesity and cardiovascular morbidity and mortality

The idea that central fat distribution is associated with higher disease risk is far from novel [5] and is supported by a wealth of epidemiological and clinical data (reviewed in [6]). In a meta-analysis of 58 prospective studies of >220,000 free of cardiovascular disease individuals, the hazard ratio (HR) for cardiovascular disease was 1.23 [95% confidence intervals (CI) 1.17–1.29] per 1 standard deviation higher baseline value of BMI, 1.27 (95% CI 1.20–1.33) with waist circumference and 1.25 (95% CI 1.19–1.31) with waist-to-hip ratio, after adjustment for age, sex, and smoking status. These associations remained significant, although attenuated, after further adjustment for systolic blood pressure, history of diabetes, and total and HDL cholesterol. Moreover, within each tertile of BMI, HRs increased log-linearly with waist circumference and waist-to-hip ratio [7]. Similar findings with hazard ratio (HR) for cardiovascular disease 1.15 [95% CI 1.04–1.27] per 1 standard deviation higher baseline value of waist circumference and 1.15 (95% CI 1.04–1.27) with waist-to-hip ratio were reported in a meta-analysis of nine British studies comprising 82,864 subjects [8].

These data illustrate the notion that measures of central fat distribution, such as waist circumference and waist-to-hip ratio, can improve risk stratification in conjunction with obesity indices. On the other hand, it should also be kept in mind that all the correlations between BMI, waist circumference and fat mass are very high (Pearson's r > 0.90), and all three show comparable correlations with VAT mass (r ≈ 0.75). Therefore, to examine the association between waist circumference and cardiovascular risk [9] and draw conclusions about its specific effect on disease risk, it is essential to adjust for BMI, which is not done in many studies.

Among individuals with established vascular disease, high BMI is sometimes associated with better prognosis, the so-called “obesity paradox” [10]. Central obesity though remains detrimental and associated with increased mortality: among 15,923 subjects with coronary artery disease (CAD) followed up for 2.3 years, HR for death was 1.70 (95% CI 1.58-1.83) after adjustment for age, sex, smoking, diabetes, hypertension, heart failure, and BMI [11]. Interestingly, when patients with CAD were categorized into four groups based on fat distribution and BMI (normal BMI/peripheral fat distribution, normal BMI/central fat distribution, obese/peripheral fat distribution, obese/central fat distribution) those with central obesity and normal BMI showed the poorest survival compared with all other groups [12].

Peripheral fat accumulation is protective

Recent systematic reviews identified and summarized data on 37,725 individuals followed-up on average for 11.2 years in nine prospective cohort studies that examine the relationship between hip circumference and cardiovascular disease risk (three studies), cardiovascular mortality (four) or total mortality (two studies) [13, 14]. All, but one, report an inverse association between hip circumference and the respective outcome in both male and female subjects. This association became apparent only after adjustment for waist circumference and/or BMI. Due to the strong bivariate associations between BMI and waist and hip circumference, as they all reflect total adiposity, it is important to adjust for BMI and waist circumference in any model to reveal the independent effects of hip circumference on cardiovascular morbidity and mortality [13, 14]. In addition, peripheral fat distribution is associated with traditional risk factors, lower blood pressure, improved insulin sensitivity and a beneficial lipid profile, as reviewed before [15].

Peripheral fat distribution is also an integral part of the MHO phenotype [16-18]. The incidence of cardiovascular disease among these individuals is in some studies comparable to the normal-weight population [17-21], and in other studies intermediate between normal-weight and metabolically unhealthy obese subjects [22], but altogether significantly lower than what would be expected solely on the basis of BMI.

Anthropometric indices for clinical predictions

The question that arises from the epidemiological studies briefly discussed above is whether measurements of fat distribution should become standard part of the every day assessment of cardiovascular risk. This question was formally addressed in the Emerging Risk Factors Collaboration study which showed that addition of BMI, waist circumference, waist-to-hip ratio or of their combination does not improve cardiovascular risk prediction over and above the one based on systolic blood pressure, history of diabetes and lipids [7]. Nevertheless, the American Heart Association in its most recent scientific statement on the assessment of adiposity recommends measurement of both BMI and waist circumference in everyday clinical practice. It also emphasizes the need for sex-, age-, race- and BMI-specific cut-off values of waist circumference [23]. In a population study, the optimal BMI and waist circumference thresholds for the identification of increased cardiovascular risk appear to be comparable in white and African American men (29.1 versus 30.4 kg/m², and 99.1 versus 99.4 cm, respectively), but significantly lower in white compared to African American women (30.0 versus 32.9 kg/m², and 91.9 versus 96.8 cm) [24]. When VAT area was assessed by computed tomography (CT) though, the threshold for increased cardiovascular risk was higher in both white men and women compared to the African American ones (140 versus 82 cm² in men 141 versus 97 cm² in women) [25]. In addition, a number of new indices have been proposed as superior to the use of BMI or of circumference measurements (summarized in Table 1).

Table 1.

New anthropometric indices.

Author	Index	Formula	Population	Predictive of	Related citations
Amato et al [26]	Visceral Adiposity Index	Males: $\left[\frac{WC}{39.68+(.88\times BMI)}\right]\times$ Females: $\left[\frac{WC}{36.58+(.89\times BMI)}\right]\times$	N=1,498 primary care subjects	• Cardiovascular events •Cerebrovascular events	Knowles et al [27] Mohammadreza et al [28] Al-Daghri et al [29] Zhang et al [30]
Arsenault et al [31]	Hypertriglyceridemic waist	Males : WC ≥ 90 cm and TG ≥ 2.0 mmol/L Females: WC ≥ 85 cm and TG ≥ 1.5 mmol/L	N=21,787 Age 45-79 y Follow-up 9.8 y	• Coronary artery disease	De Graaf et al [32] Blackburn et al, [33] Zhang et al [30]
Bergman et al [34]	Body Adiposity Index	$\frac{HC}{Heigh{t}^{1.5}}-18$	N=1,733 Mexican Americans Age 35 (18-67) y	• % body adiposity	Melmer et al [35] Moliner-Urdiales et al [36]
Ashwell et al [37]	Waist-to-Height Ratio	Waist/Height	N=305,851 (meta-analysis)	• Cardiovascular risk factors • Myocardial infarction Cardiovascular mortality	Sluik et al [38]
Trefethen LN [39]	“New BMI”	$\frac{1.3\times Weight}{Heigh{t}^{2.5}}$	No data available		http://people.maths.ox.ac.uk/trefethen/bmi.html (accessed 03/25/13)

BMI: Body Mass Index; HDL: High Density Lipoprotein; TG: Triglycerides; WC: Waist Circumference.

Insights from imaging studies

From a different point view, it can be argued that anthropometric indices are only indirect measures and possibly poor approximations of the respective adipose tissue depots. Waist circumference reflects VAT as well as abdominal SAT, while hip circumference depends on the amount of gluteal fat, but also on muscle mass. With the advent of detailed imaging techniques, it is possible to obtain more accurate measurements of the above depots, together with smaller, so far overlooked, depots.

CT and magnetic resonance imaging (MRI) studies can examine the independent associations of abdominal SAT and VAT with risk factors and disease. Both earlier [40] and more recent studies indicate that VAT [41, 42] and the ratio between VAT/SAT [43] show the closest correlations to cardiometabolic risk. In addition, increased VAT for comparable BMI levels appears to mediate the adverse metabolic profile of South and East Asian populations [44, 45].

Epicardial and perivascular adipose tissues in relation to atherosclerosis

Given their anatomical proximity to the heart, the coronaries, or other susceptible to atherosclerosis arteries, the epicardial, pericardial and perivascular adipose tissue depots have attracted considerable interest over the last few years. In cross-sectional studies, higher amounts of epicardial/pericardial adipose tissue are associated with higher prevalence of cardiovascular disease; this relationship is seen in both sexes and is attenuated after adjustment for BMI and other risk factors [46-48]. Interestingly, increased pericardial and peri-aortic depots do not always coexist and significant discordance is noted in more than 20% of individuals [49]. The thoracic peri-aortic adipose tissue is more closely associated to central obesity and metabolic syndrome, but its associations with cardiovascular risk or disease are less significant and become apparent only after adjustment for VAT [48-50].

More recently, prospective data have started accumulating: the volume of epicardial adipose tissue [51] and its expansion [52] were both associated with an increase of coronary artery calcification over time. In patients with established coronary artery disease (n=194, age 59.4 years, 80% men, BMI 28.7 ± 4.6 kg/m²), EAT measured by echocardiography was not predictive of new major cardiovascular events over a follow-up period of 3.6 years [53]. In the general population though, EAT is predictive of future coronary events: among 4,093 participants in the Heinz Nixdorf Recall cohort study (age 59.4 years, 47% men, follow-up 8.0 years), doubling of EAT was associated with a 1.5-fold risk of coronary events after adjustment for cardiovascular risk factors [HR 1.54 (95% CI 1.09 – 2.19)][54].

Other small depots: kidney and breast

Renal sinus fat, a small depot (volume of 4.6±3.2 cm³ in MRI) is of interest because it surrounds directly the renal artery and vein. Increased renal sinus fat, also called “fatty kidney”, was associated with the presence of hypertension in two middle-aged populations and this relationship remained significant after accounting for BMI and/or VAT [55, 56]. The breast adipose tissue has been studied in MRI studies of healthy, premenopausal women with a wide range of adiposity and its volume correlated positively with central obesity and VAT, and negatively with leg adipose tissue. No associations were found with cardiovascular risk factors in either of these studies [57, 58].

Pathophysiological mechanisms that link specific depots to cardiometabolic disease

Elucidating the mechanisms that differentiate the significance of each depot is a challenging scientific question that is potentially important for disease prevention and treatment. Factors like the cellular characteristics, the location and the function of different adipose tissue deposits have all been linked to their effects in three major hypotheses, as discussed in the following paragraphs (see also Figure 1).

Figure 1. Adipose tissue depots influence the development of atherosclerosis through multiple direct and indirect mechanisms.

Subcutaneous adipose tissue depots (abdominal, gluteal and femoral) release in the systemic circulation free fatty acids (FFA) and numerous adipokines/cytokines that affect the function of critical tissues (liver, skeletal muscle, kidney, heart, vessel wall) and have detrimental effects on traditional (dyslipidemia, hypertension, diabetes) or novel (inflammation) cardiovascular risk factors. Visceral adipose tissue products are drained through the portal vein directly to the liver, affect hepatic function and a percentage of them enters the systemic circulation and has systemic effects. Smaller depots may have local effects: the perivascular fat can interact with the vessel wall, the epicardial fat with the myocardium and the coronary arteries and the renal sinus fat with the renal artery/vein.

A. The subcutaneous sink and the toxicity of ectopic adipose tissue

The subcutaneous depots are often considered as the main adipose sites that ensure effective energy deposition under conditions of caloric excess and release under conditions of negative energy balance. According to this hypothesis, accumulation of fat as VAT and in the smaller internal depots reflects the inability of SAT for further triglyceride storage and has detrimental effects on the surrounding tissues, particularly liver, skeletal muscle and the heart [59]. This is evident in the case of lipodystrophies [60] and in the use of thiazolidinediones (rosiglitazone or pioglitazone): administration of pioglitazone to overweight/obese, insulin resistant but non-diabetic subjects induced adipogenesis and expansion of SAT, reduction of VAT and improved insulin sensitivity [61].

B. Adipokines and endocrine effects

By synthesizing and releasing a number of cytokines and other proteins, adipose tissue exerts endocrine effects on a number of tissues, the vasculature itself and others that contribute indirectly to cardiovascular disease, like liver and skeletal muscle. Adipokines like, leptin, interleukin-6 and -8, monocyte chemoattractant protein 1, serum amyloid A, plasminogen activator inhibitor 1, angiotensinogen and an array of others, are upregulated in obesity and can affect all stages of atherosclerosis from monocyte recruitment to foam cell formation, smooth muscle cell proliferation, plaque destabilization and thrombosis (reviewed in [62, 63]).

Although a detailed description of their individual effects is outside of the scope of this review, specific mention is required of adiponectin, a highly abundant circulating hormone. In contrast to other adipokines that are increased in obesity, adiponectin is reduced in obese individuals. In addition to BMI, adiponectin levels are influenced by sex (higher in women compared to men), ethnicity (higher in Caucasians compared to Hispanics and African Americans) and fat distribution (positively correlated with lower body fat and negatively with truncal fat) [64]. Circulating adiponectin levels are associated with cardiovascular disease, but in a complex way: in the general population, adiponectin is inversely related to traditional risk factors and future events; in older individuals, this association is U-shaped with a protective effect for concentrations up to 12.4 mg/L [65]. On the contrary, in populations with pre-existing cardiovascular disease, higher adiponectin levels are associated with poorer prognosis [65-67]. In agreement with epidemiological data, recent in vitro studies suggest that under certain experimental conditions adiponectin can have adverse effects on macrophages, T cells [68], in addition to its better established atheroprotective effects (reviewed in [69, 70]). Thus, the exact role of adiponectin in chronic inflammatory states remains for now poorly understood [71].

C. The portal hypothesis and other local/direct effects of smaller depots

The portal hypothesis posits that, when enlarged, the lipolytically active VAT releases significant amounts of free fatty acids directly to the portal circulation and consequently to the liver [72]. These in turn will be transported to the systemic circulation in very-low-density lipoproteins, will stimulate hepatic glucose production and suppress hepatic insulin clearance, leading to dyslipidemia, hyperinsulinemia and insulin resistance [72]. Indeed, it was later shown that with visceral obesity the contribution of VAT to the free fatty acids that reach the liver increases from <10% to over 30-40% [73]. However, increased concentration of free fatty acids in the portal compared to the systemic circulation has not been documented in humans (reviewed in [74]) and the majority of free fatty acids that reach the liver is derived from SAT, and particularly the upper body SAT even in obese individuals [73]. In addition to free fatty acids, VAT releases interleukin-6 to the portal vein, which in turn induces synthesis of C-reactive protein in the liver and contributes to systemic inflammation [75].

Smaller adipose tissue depots, like the epicardial one, are probably devoid of systemic effects. On the other hand, they may have pathophysiological significance because of their effects on the heart and the coronary arteries. In in vitro studies, adipokines produced by EAT can induce expression of adhesion molecules in human coronary artery endothelial cells and migration of monocytes [76] and affect the contractile function and insulin responses of cardiomyocytes [77], as well as fibrosis of the atrial myocardium [78]. Perivascular adipose tissue can regulate vascular tone and thus blood pressure control through the release of a relaxing factor, recently identified as methyl-palmitate [79]. Finally, obesity is accompanied by reduced angiogenesis and capillarization within the adipose tissue itself, which may be linked to the concomitant inflammation and insulin resistance [80, 81].

Experimental studies to firmly establish the independent roles of separate depots

It has to be noted that although new data accumulate steadily on the characteristics of each adipose tissue depot and on their relationship to metabolic and cardiovascular diseases, studies that will establish a direct, causal relationship are extremely difficult. They require experimental manipulations that will target a specific depot and alter its function without affecting the rest of the depots or total body adiposity. Very few studies provide such information, but there are some important insights from some clinical studies and animal models, as described below.

A mouse model without perivascular adipose tissue

Murine perivascular adipose tissue has features of brown adipose tissue and as a result, it can generate heat to preserve intravascular temperature upon cold exposure [82]. A mouse model in which the adipogenic transcription factor PPARγ was deleted specifically in smooth muscle cells (SMPG KO mice) was characterized by complete lack of perivascular adipose tissue with normal subcutaneous and gonadal depots. Studies of this model show that activation of perivascular fat by cold exposure can attenuate high-fat diet induced atherosclerosis, at least partly through enhanced lipid clearance and prostacyclin release [82]. Although the human relevance of these findings is unclear, the study demonstrates the potential of this depot to modulate vascular damage in vivo.

Omental adipose tissue in humans: how important is it?

As already discussed, the links between VAT and cardiometabolic disease have attracted extensive interest. However, VAT correlates closely with intrahepatic TG content and it has been proposed that the former represents mostly a marker of the latter, rather than a hazardous tissue per se. Thus obese, middle-aged individuals with high intrahepatic TG content have reduced insulin sensitivity in comparison to others with equal VAT but low intrahepatic TG. On the contrary, individuals with high VAT were as insulin sensitive as those with low VAT provided that they were matched on intrahepatic TG content [83]. These findings are in accordance with interventional studies that involve removal of the omental depot in obese individuals. In eight out of nine published studies, addition of omentectomy to gastric bypass or sleeve gastrectomy did not augment weight loss or the improvement in insulin sensitivity and other metabolic indices achieved by the bariatric intervention itself (Table 2) [84-91]. Although longer-term follow-up and the incidence of cardiovascular events can be assessed in future studies, these data question the causative role of omental adipose tissue in metabolic dysfunction.

Table 2.

Omentectomy, weight loss and metabolic parameters.

Author	N	Subject characteristics	Treatment	Follow-up	Outcome
Fabbrini et al [84]	22	Obese men and women	Roux-en-Y gastric bypass ± omentectomy	12 months	Weight loss, lipids, FPG, insulin sensitivity: no difference between groups
Fabbrini et al [84]	7	Obese men and women with T2DM	Omentectomy	3 months	Insulin sensitivity, diabetes medications: no change compared to baseline
Herrera et al [85]	22	Obese men and women	Roux-en-Y gastric bypass ± omentectomy	1, 3, 6, 12 months	Weight loss, lipids, FPG, insulin, adipokines: no difference between groups
Dillard et al [86]	28	Obese men and women	Roux-en-Y gastric bypass ± omentectomy	3 months	Weight loss: no difference between groups FPG, total- and VLDL-cholesterol: reduction only in the omentectomy group
Tamboli et al [87]	21	Obese men and women	Roux-en-Y gastric bypass ± omentectomy	6, 12 months	Skeletal muscle gene expression: greater reduction of inflammatory genes in the omentectomy group
Wu et al [88]	40	Obese men and women	Sleeve gastrectomy ± partial enterectomy and omentectomy	1, 3, 6, 12 months	Weight loss: no difference between groups
Dunn et al [89]	40	Obese men and women	Roux-en-Y gastric bypass ± omentectomy	1 month	Weight loss, insulin sensitivity, hepatic glucose production, hepatic insulin sensitivity index : no difference between groups
Lima et al [90]	20	Obese premenopausal women with MS	Roux-en-Y gastric bypass ± omentectomy	1, 6, 12 months	Weight loss and reduction in CRP: greater in omentectomy group. Insulin sensitivity: no difference between groups
Sdralis et al [91]	21	Obese men and women	Sleeve gastrectomy ± omentectomy	7 days, 1, 3, 12 months	Weight loss, lipids, FPG, insulin, CRP, adipokines: no difference between groups

CRP: C-reactive protein; FPG: fasting plasma glucose; T2DM: type 2 diabetes mellitus; VLDL: very low density lipoprotein

Diet, exercise and visceral adiposity

Identification of dietary patterns associated preferentially with central obesity and high cardiovascular risk can be of major public health significance. In a cross-sectional assessment of diet of 497,308 individuals (age 25-70 years, 70.7% women) from 10 European countries, adherence to a Mediterranean diet, enriched in vegetables, fruits, legumes and unsaturated fatty acids, low in meat and dairy products, was associated with significantly lower waist circumference in both men and women, but not with BMI [92]. A similar dietary pattern, high in fruits and low in processed meat, margarine, soft drinks and white bread was related to smaller increases in waist circumference during 5.5 years of follow-up in the 48,631 participants of the European Prospective Investigation into Cancer and Nutrition study [93]. These observational data support the findings of a recent randomized trial: a study comparing two Mediterranean diets, supplemented with olive oil or mixed nuts, with a control, low fat, diet was terminated prematurely due to lower incidence of major cardiovascular events in the Mediterranean diet groups (multivariable adjusted hazard ratio 0.70 (95% CI 0.54 – 0.92) [94].

Caloric restriction induces weight loss and women appear to lose more VAT as percent of total body fat loss compared to men [95]. However, there is no specific diet that preferentially targets VAT: diets with 25% versus 15% protein, 40% versus 20% fat, 65% versus 35% carbohydrates, all resulted in similar reductions in total body fat, lean mass, abdominal SAT and VAT [95].

Exercise is the second feature of healthy lifestyle. In overweight and obese adults, aerobic exercise alone, without concomitant caloric restriction, can lead to significant reduction of VAT after 12 weeks; on the contrary, resistance training has no significant effects on VAT [96, 97]. In addition, some studies report that exercise can facilitate VAT reductions during diet-induced weight loss [98] or prevent VAT regain after successful dieting and weight loss [99] (and reviewed in [6]).

Brown adipose tissue in humans: it exists – how important is it?

The “rediscovery” of brown adipose tissue (BAT) in adult humans (100) has created a lot of scientific interest on the molecular and physiological properties of this depot. Based on its ability to dissipate energy as heat by uncoupling oxidative phosphorylation and ATP synthesis (a function of uncoupling protein-1, UCP-1), BAT is seen as a potential new therapeutic tools to combat obesity. Although outside of the scope of this review, there is now evidence that UCP-1 expressing adipocytes are present in humans in small depots with the characteristics of classical BAT (101), but also interspersed in the WAT depots, so-called “brite” or “beige” adipocytes (102). The amount of BAT depends on several factors including sex, age and BMI (103) and is increased after surgical weight loss (104). In animal models, multiple hormones, molecular pathways and pharmacological agents (from catecholamines, natriuretic peptides, the retinoic acid pathway to irisin and thiazolidinediones, reviewed in 105) can expand and activate brown/brite-beige fat. Translational studies will be needed to clarify whether these tissues affect whole body metabolism in humans and to identify efficient tools to employ them towards cardiometabolic health.

Conclusions

Obesity is now recognized as a major public health issue and a driver of a number of comorbidities, including cardiovascular disease. A more recent notion is that obesity is a heterogeneous condition. Body fat distribution, which reflects the size of the larger fat depots, provides an additional level of risk classification for obese individuals and is now recommended part of every day clinical practice. Smaller depots, like the epicardial, when located in critical areas, may interact directly with important tissues, like the vessel wall, the myocardium and the renal vessels, but their in vivo significance remains unknown. Although epidemiological studies have provided and continue to provide ample data, unraveling the effects of each depot in experimental studies remains a big challenge. As a result, therapeutic approaches that will target specific fat deposits are currently lacking and maintaining a healthy body weight through lifestyle measures is still the cornerstone of cardiometabolic health.