Mapping body fat distribution: A key step towards the identification of the vulnerable patient?

Abstract

Although excess body fat is a significant health hazard, estimation of body fat content with the body mass index may not adequately reflect the amount of atherogenic adipose tissue (AT), i.e. visceral and ectopic fat. As opposed to subcutaneous AT that supposedly acts as a metabolic sink buffering excess dietary energy, visceral or intra-abdominal AT depots respond to several external stimuli that trigger lipolysis and secretion of free fatty acids (FFAs). Reaching the liver, FFAs accumulate in the liver and, over time, promote a chronic condition known as non-alcoholic fatty liver disease (NAFLD). The liver of the typical NAFLD patient secretes large amounts of very-low-density lipoproteins, the lipid content of which may accumulate in additional organs (skeletal muscle, heart, and pancreas). Here, we review the evidence emerging from functional and population studies that point towards an important role of ectopic fat accumulation in the pathophysiology of type 2 diabetes and cardiovascular disease. We conclude that although patients with impaired glycemic control or type 2 diabetes are at increased cardiovascular disease (CVD) risk, estimating cardiovascular risk goes wellbeyond the assessment of glycemic control and traditional CVD risk factors, and the estimation of visceral/ectopic fat deposition via readily available imaging techniquesshould be considered.

Key words::

• Atherosclerosis
• computed tomography
• ectopic fat accumulation
• glucose–insulin homeostasis
• imaging
• inflammation
• insulin resistance
• magnetic resonance imaging
• type 2 diabetes
• visceral obesity

Abbreviations
BMI	=	body mass index
CAC	=	coronary artery calcium
CHD	=	coronary heart disease
CIMT	=	carotid artery intima–media thickness
CRP	=	C-reactive protein
CT	=	computed tomography
CVD	=	cardiovascular disease
EAT	=	epicardial adipose tissue
FFA	=	freefatty acid
FHBL	=	hypobetalipoproteinemia
HDL	=	high-density lipoprotein
IHD	=	ischemic heart disease
IL-6	=	interleukin-6
IMCL	=	intramyocellular lipid
LDL	=	low-density lipoprotein
MRI	=	magnetic resonance imaging
NAFLD	=	non-alcoholic fatty liver disease
NASH	=	non-alcoholic steatohepatitis
NMR	=	nuclear magnetic resonance
PAT	=	pericardial adipose tissue
PPAR-γ	=	peroxisome proliferator-activated receptor-γ
SAT	=	subcutaneous adipose tissue
VAT	=	visceral adipose tissue

Key messages

• Independent of total body fat, visceral (intra-abdominal) and hepatic fat accumulation are significant predictors of a deteriorated cardiometabolic risk profile and of an increased risk for type 2 diabetes and cardiovascular disease.

• Imaging techniques are now available to estimate the amount of ectopic fat accumulation in the intra-abdominal adipose depots, the liver, the heart, and the skeletal muscle.

• Although patients with type 2 diabetes are characterized by an increased cardiovascular disease (CVD) risk, a better ‘phenotyping’ of these patients would require going beyond the assessment of glycemic control and traditional CVD risk factors and rather measure or at least estimate the extent of ectopic fat accumulation.

Introduction

It is now widely accepted that obesity causes prejudice to several aspects of global human health (1–7). However, some individuals carrying what could be considered as a small excess of body fat may nevertheless show early signs of insulin resistance and substantial deterioration in their plasma lipoprotein-lipid profile, which are predictive of increased diabetes and cardiovascular disease (CVD) risk (8). Assessing high-risk obesity in clinical practice represents a challenge to physicians and health care professionals due to its remarkable heterogeneity. Prospective and cross-sectional imaging studies published over the last 20–25 years have provided robust evidence that the regional distribution of body fat, rather than excess fatness per se, is the critical correlate of the metabolic abnormalities expected from the presence of overweight/obesity. Although waist circumference should be measured by all means beyond the body mass index (BMI) to estimate the amount of atherogenic abdominal adipose tissue accumulation, the assessment of waist circumference does not discriminate between visceral (VAT) versus subcutaneous adipose tissue (SAT) accumulation, the latter fat depot being associated with an overall more favorable cardiometabolic risk profile (9). It has been proposed that VAT accumulation could represent a good marker for the presence of more fundamental neuroendocrine abnormalities leading to storage of lipids at undesired sites, such as the liver, the heart, the skeletal muscle, and the pancreas, a phenomenon described in the literature as ‘ectopic fat’ deposition (10). The main objective of this paper is therefore to provide an overview of the relationship between visceral fat, liver fat, heart fat, and muscle lipids and metabolic abnormalities predisposing individuals with such ectopic fat accumulation to develop premature CVD and/or type 2 diabetes.

Abdominal fat and visceral adipose tissue

In 1947, a French physician named Jean Vague was the first to propose that excess upper body fat (a phenotype frequently found in men) was more closely associated with greater cardiovascular complications than other forms of obesity such as those often found in premenopausal women who tend to accumulate lower body fat in the gluteal-femoral areas (11). In fact, it is well known that premenopausal women must accumulate a substantial amount of body fat in order to develop the metabolic abnormalities found in men.

The epidemiological evidence

Visceral fat represents approximately 10% of total body fat in women whereas this proportion may reach up to 20%–25% in men (12). As abdominal fat accumulation is modulated by steroid hormones, postmenopausal women tend to develop a pattern of visceral fat accumulation which eventually approaches over years to the one found in men, a phenomenon which considerably raises their CVD and type 2 diabetes risk (13,14). On top of gender and steroid hormones, several other biological factors explain the regional distribution of abdominal fat such as age, ethnicity (for a given waist circumference, Asians have more visceral fat than Caucasians, who in turn have more visceral fat than blacks (15–17)), stress, lack of physical activity/exercise, diet, and genetic factors (18,19). Although it is beyond the scope of this review article to summarize the biological mechanisms that lead to increased visceral and ectopic fat accumulation (see other timely reviews on the topic (20–22)), we will briefly describe the evidence that points towards a detrimental role of visceral adiposity and ectopic fat deposition in the risk of type 2 diabetes and CVD.

Over the past decade, a few studies have investigated the relationship between VAT and incident coronary heart disease (CHD). In a prospective study including 291 men who were followed for 2.2 years, Kuk and colleagues (23) have shown that the amount of VAT was an independent predictor of all-cause mortality. Importantly, the relationship between VAT and mortality was independent of other indices of body fat distribution such as SAT, liver fat, as well as waist circumference. In a sample of Japanese men followed for 10 years, those who incidentally developed CHD during follow-up had comparable BMIs as well as similar total and SAT accumulation, whereas CHD cases had approximately 22% more VAT than controls (24). VAT was also found to be an independent predictor of incident type 2 diabetes in that cohort (25). In women aged 70–79 years in the Health, Aging and Body Composition Study, VAT was an independent predictor of incident myocardial infarction over a 4.6-year follow-up (26), the relationship between VAT and incident myocardial infarction being independent of BMI and total fat mass. Among men of the same age-range who were included in the study, there was no relationship between any of the studied adiposity markers and incident myocardial infarction. This sex-specific relationship may be attributable to a survival bias and/or to the fact that with advancing age, increased body fat accumulation may represent a ‘protective’ fuel source. However, both VAT and SAT were associated with the risk of chronic heart failure in the Health, Aging and Body Composition Study (27). In the Multicultural Community Health Assessment Trial (M-CHAT), Lear and colleagues(28) tested the potential cross-sectional associations between VAT and carotid artery intima–media thickness (CIMT) measured by carotid ultrasound in apparently healthy middle-aged individuals. Among the 794 participants of the study, they found that VAT was positively associated with CIMT and with total plaque area even after adjusting for potential confounders including age, ethnicity, smoking, and total body fat. In men, but not in women, VAT was associated with CIMT after further adjustment for glucose and insulin levels, homocysteine levels, blood pressure, the total cholesterol to high-density lipoprotein (HDL) ratio, low-density lipoprotein (LDL) cholesterol, triglycerides, apolipoprotein B levels, and waist circumference. Finally, in an older cohort of subjects without documented CHD, the Rancho Bernardo Study, Kramer and colleagues (29) showed that abdominal obesity was an independent predictor of coronary artery calcium (CAC) progression in both sexes. In this sample of 156 men and 182 women (mean age of 67 years), results showed that waist circumference was the best predictor of CAC progression in women, whereas the ratio of VAT to SAT was the most powerful predictor in men.

Visceral adipose tissue and the metabolic syndrome

Aiming at identifying the potential biological mechanisms that may explain the association between VAT and CVD/type 2 diabetes risk, several investigators have suggested that increased VAT is the key correlate of the metabolic syndrome (30–32). The metabolic syndrome represents a constellation of CVD risk factors including abdominal obesity, hypertension, a deteriorated lipoprotein-lipid profile, a pro-inflammatory and pro-thrombotic profile, as well as insulin resistance, leading to glucose intolerance and type 2 diabetes (33,34). Meta-analyses conducted on the topic have shown that the presence of the metabolic syndrome increased the risk of CVD approximately 1.5–2.0-fold (35,36). Although the etiology of the metabolic syndrome is still a matter of debate (37), cross-sectional and prospective studies have shown that the relationship between VAT and metabolic syndrome was independent of other body fat compartments such as SAT, specific measures of insulin resistance, or cardiorespiratory fitness (38–42). A substantial number of cross-sectional studies have also shown that independently of anthropometric indices such as BMI, waist circumference, and SAT, the amount of VAT was positively associated with plasma levels of C-reactive protein (CRP), a marker of inflammation, which has emerged as a significant correlate of CVD risk (43–45).

Subcutaneous adipose tissue: a protective body fat compartment?

Although excess abdominal fat is detrimental to human health, experimental and imaging studies have provided evidence that irrespective of VAT accumulation, some SAT depots may represent metabolic sinks buffering the energy surplus towards potentially protective body fat compartments. Among those pioneer studies, investigators of the Framingham Heart Study have recently provided convincing evidence for a cardioprotective role of increased SAT accumulation (46). In that study, although both SAT and VAT were positively associated with a deteriorated cardiometabolic risk profile, a thorough classification of the study participants on the basis of both VAT and SAT accumulation revealed that among participants with the highest VAT accumulation, those with the highest SAT accumulation were characterized by the lowest triglyceride levels. Among men with the highest VAT accumulation, higher SAT levels were also associated with a lower prevalence of low HDL cholesterol levels (46). Although this study was limited by its cross-sectional design, it provided evidence for a cardioprotective role of SAT. In the same study population, VAT was found to be a better correlate of specific features of insulin resistance compared to SAT (47). These results, combined with those of epidemiological studies which have suggested a cardioprotective role of carrying an elevated hip circumference (an anthropometric marker of gluteo-femoral fat) for a given waist circumference, further reinforce the notion that there is a need to go beyond traditional indices of body fat distribution to appreciate better the role of regional adipose tissue distribution as a potential risk factor for cardiovascular mortality and morbidity (6,48). However, beyond the cardioprotective role of SAT, one must not disregard the direct impact of an increased VAT on cardiovascular health.

The biological evidence

In order to appreciate the deleterious consequences of visceral fat on various metabolic pathways, a closer look at the pathobiology of visceral adipocytes is warranted. In lean and healthy individuals, insulin has a critical role to play in maintaining optimal adipocyte function (differentiation, lipid storage, lipolysis, etc.). As VAT mass expands, as a consequence of overeating, lack of physical activity, insulin resistance, and inflammation, insulin is no longer capable of controlling adipocyte differentiation and the balance between lipid uptake and lipolysis. Despite not being able to divide, visceral adipocytes continue to remove lipids from the blood-stream and therefore become hypertrophied (49). In parallel, a local state of hypoxia controls macrophage infiltration inside VAT, which further preventsperoxisome proliferator-activated receptor(PPAR)-γ-dependent fat cell differentiation through inflammatory mediators such as tumor necrosis factor-α or interleukin (IL)-6 (50,51). Interestingly, IL-6 also has a key role to play in promoting hepatic C-reactive protein secretion, another detrimental consequence of the portal vein system draining fat cells in the visceral adipose tissue (52). Investigators who have directly studied the anatomical characteristics of abdominal adipose tissue have reached the conclusion that visceral adipocytes, e.g. those located in the omentum and around the mesentery which are drained by the portal vein, directly carry the VAT-derived secretagogues to the liver, whereas adipose cells of subcutaneous depots are drained by the peripheral vein system (53). According to this ‘portal vein theory’, freefatty acids (FFAs), which are secreted by visceral adipocytes in a dose-dependent manner (increased VAT correlated with portal FFA levels), are directly conveyed to the liver where they accumulate. This phenomenon under which FFAs and triglycerides accumulate inside the liver is an important cause of liver fat infiltration, which may lead to non-alcoholic fatty liver disease(NAFLD). In a heterogeneous non-diabetic population, Thamer and colleagues (56) have shown that even after adjusting for sex, age, waist-to-hip ratio, and SAT, VAT was a powerful predictor of intrahepatic lipid accumulation as measured by proton magnetic resonance spectroscopy. Other groups have also reported similar observations and further suggested that visceral fat changes over time may be associated with the onset of NAFLD (57,58).

Visceral fat and inflammation

Among the other potential links between excess visceral adiposity and hepatic fat accumulation, non-infectious or systemic inflammation is also recognized as a key player. It has been documented for decades that inflammation plays an important role in the pathophysiology of several chronic diseases such as CVD and type 2 diabetes (61–63). However, recent data point towards VAT as an important source of pro-inflammatory and pro-thrombotic adipokines that may promote hepatic and peripheral insulin resistance and favor plaque occlusion and rupture (64,65). In fact, we now know that several inflammatory cells such as leukocytes and macrophages infiltrate and reside within VAT where, upon activation, they create a local state of inflammation and insulin resistance (66,67). Among the inflammatory markers secreted by tissues localized inside VAT, IL-6 may contribute to the local insulin resistance state inside adipose tissue as well as in the liver by reducing the expression of insulin receptor substrate-1, a key protein involved in insulin signaling (68,69). It has recently been shown that VAT accumulation was an independent predictor of circulating IL-6 concentrations (70). A study also showed that portal vein IL-6 concentrations were twice as high as IL-6 concentrations inside the brachial artery (71). As portal vein IL-6 concentrations were also closely associated with brachial artery CRP concentrations, this study provided strong experimental evidence that VAT is a sensible predictor of systemic inflammation. Accordingly, results of the Québec Cardiovascular Study have shown that the long-term relationship between IL-6 and risk of ischemic heart disease (IHD) was independent of plasma CRP levels, whereas the opposite was not found, i.e. the relationship of CRP to IHD was not independent of IL-6 (72). Similar results have also recently been published in a sample of middle-aged men living in England (73). On top of promoting hepatic CRP production, IL-6 increases insulin resistance in hepatocytes and stimulates hepatic gluconeogenesis (69,74). In contrast to IL-6, an expanded VAT area has been associated with low plasma levels of adiponectin, an anti-inflammatory and anti-diabetic adipokine (75). Adiponectin is specifically secreted by adipocytes. In normal-weight individuals, adiponectin is one of the most abundant proteins in the blood-stream that might protect the vascular endothelium from lipid and monocyte infiltration (76). Several investigations have also reported that adiponectin has an important role to play in hepatic lipoprotein-lipid metabolism (77,78). For instance, in a murine model of NAFLD, injection of recombinant adiponectin has been shown to lower hepatic steatosis, hepatic inflammation, as well as serum transaminase concentrations (79). The next section will discuss the biology and the pathological consequences of hepatic fat accumulation.

Liver fat

The liver is the central organ of the lipoprotein-lipid metabolism. In genetically predisposed individuals as well as in many who adopt poor life-style habits, small droplets of triglyceride may accumulate in the liver. Accumulation of hepatic fat may eventually lead to NAFLD, which encompasses the spectrum that links the fatty liver, non-alcoholic steatohepatitis (NASH), and eventually liver cirrhosis. Under homeostatic conditions, the relative percentage of fat in the liver is below 5% of its total weight (80). NAFLD is defined as the proportion of fat accumulation in the liver exceeding 5%–10% of total liver weight, as determined from the percentage of fat-laden hepatocytes by light microscopy (81). It has been shown that the worldwide prevalence of fatty liver is directly correlated with the prevalence of obesity (82). For instance, in the US, the prevalence of obesity (defined as BMI ≥30 kg/m²) is identical to the prevalence of NAFLD, reaching an astonishing 30% of the general population (83).

Several studies have shown that liver fat is closely associated with the components as well as with the development of the metabolic syndrome. Recently, Kotronen and colleagues have provided strong evidence supporting the idea that liver fat is a hall-mark of type 2 diabetes (84). After having carefully matched patients with type 2 diabetes with non-diabetic patients for age and anthropometric indices such as BMI or waist circumference, patients with type 2 diabetes had 80% more liver fat than non-diabetic patients, while showing 16% more visceral fat. However, because of the cross-sectional design of this study, a causal relationship between liver fat and type 2 diabetes could not be established. However, one could suggest from these results that liver fat should also be considered as a prevalent marker/feature of type 2 diabetes. Another study conducted among patients with type 2 diabetes has revealed that insulin resistance, dyslipidemia, and systemic inflammation were only found among those who were characterized by excess liver fat (85). Therefore, the direct measurement of liver fat by proton magnetic resonance spectroscopy may be helpful as a sensitive predictor of type 2 diabetes, independent of waist circumference and/or liver enzymes. However, it is not clear whether visceral fat, liver fat, or both, should be considered to provide a sensitive prognosis of either CVD or type 2 diabetes risk. In this regard, it has been shown that both VAT and liver fat predicted a deteriorated metabolic risk profile among men, independently of other covariates such as SAT and cardiorespiratory fitness (86). In a sample of premenopausal women, VAT but not liver fat was a powerful predictor of a deteriorated metabolic risk profile (87). This discrepancy in the association of visceral and ectopic fat with cardiometabolic risk across these studies may be attributable to the low prevalence of liver fat in premenopausal women compared to men. Following menopause, however, the prevalence of NAFLD increases to a point that it can reach the levels observed in men. On the other hand, a recent study by Fracanzani et al. (88) has shown that NAFLD could be present even in individuals with a low waist circumference. However, investigators of this study estimated visceral fat by measuring waist circumference rather than obtaining a direct measurement with imaging techniques. In subjects with type 2 diabetes, VAT and liver fat (but not SAT) were both independently associated with basal hepatic insulin resistance and residual endogenous glucose production measured during the glucose-insulin clamp procedure (89). Similar to VAT accumulation, ethnic differences in liver fat accumulation can be observed. In a recent population-based study, it was found that liver fat accumulation was closely linked to visceral adiposity, regardless of ethnicity, in a large group of adults (90). However, in this study, for a given body mass, African-Americans tended to accumulate less VAT than either Caucasians or Hispanics.

From the above studies, it is difficult to determine which factor has the most important role to play in the etiology of insulin resistance. In that regard, Amaro et al. (91) have recently studied the question in patients with familial hypobetalipoproteinemia (FHBL), a genetic disorder that impairs triglyceride-rich lipoprotein secretion from the liver. Given that patients with FHBL had similar hepatic and muscle insulin sensitivity indices to lean individuals, indices that were, in both cases, considerably higher than those of individuals matched for hepatic fat content but without FHBL (patients with NAFLD), it could be suggested that liver fat in isolation cannot cause insulin resistance. However, in the setting of a dysmetabolic environment such as what can be observed in individuals with increased VAT accumulation, liver fat may associate with local and peripheral insulin resistance.

The fatty heart

In 1933, Smith and Willius (92) conducted a series of pioneering autopsy studies,the results of which suggested that obesity is frequently associated with lipid infiltration in various cavities of the heart and may therefore represent an important cause of dilated cardiomyopathy and associated CVD risk in humans. In the following decades, extensive investigations have shown that fat infiltration throughout the heart may represent an important but often forgotten cause of non-ischemic dilated cardiomyopathy in humans and a direct cause of atherosclerosis (93). There are three major fat depots in the heart: pericardial fat, epicardial fat, and intracellular fat (94). Pericardial adipose tissue (PAT) is located outside the pericardium, whereas epicardial adipose tissue (EAT) is located at the surface of the heart where it closely surrounds coronary arteries even as they penetrate the myocardium, in the absence of any discernable fascia separating EAT from the vessel wall (95). Intracellular fat is packaged in lipid droplets inside the cytoplasm of the cardiac muscle (96). Altogether, adipose tissue may cover up to 80% of the heart's surface and may encompass up to 20% of its total weight (97). Although the long-term relationship between lipid infiltration of the heart (in any depot) and CVD and/or type 2 diabetes risk is largely unknown, a few cross-sectional studies have reported a potential association between the fatty heart and CVD. For instance, Djaberi et al. (98) have measured the volume of EAT in 190 patients undergoing multi-slicecomputed tomography (CT) coronary angiography and reported an independent association between EAT and atherosclerosis burden as estimated by the coronary artery calcium score. Similar associations were reported by Gorder et al. (99), although this finding appeared to be limited to lean individuals. In another cross-sectional study, EAT thickness as well as the extent of coronary atherosclerosis measured by quantitative coronary angiography were obtained in 527 patients undergoing their first coronary angiographic procedure (100). Results of this study showed that EAT was close to 3-fold thicker in patients with documented coronary artery disease compared to those without. Results of this study also showed that EAT was more closely associated with VAT compared to either BMI or waist circumference. Additional studies have reported significant associations between pericardial fat and health hazards. For instance, using dual source CT angiography to measure simultaneously PAT and coronary atherosclerosis burden in 286 consecutive patients, Greif et al. (101) have recently shown that patients without plaques had low PAT whereas patients with any kind of plaques (calcified, non-calcified, or a mixture of all plaques) had approximately twice as much PAT. In that study, compared to well established CVD risk factors such as smoking, arterial hypertension, hypercholesterolemia, diabetes mellitus, and an elevated Framingham risk score, PAT was by far the most important risk factor for coronary atherosclerosis. For instance, compared with patients with a pericardial fat volume lower than 300 cm³, those with a pericardial fat volume above 300 cm³ had a relative risk for coronary atherosclerosis of 4.1 (95% CI 3.6–4.3). It has also been suggested that PAT was positively associated to the extent of coronary atherosclerosis in African-Americans with type 2 diabetes and that ethnic differences in PAT accumulation may explain to a certain extent the ethnic differences in susceptibility to atherosclerosis (102).

The relationship between PAT, VAT, and CAC was also investigated in 1,155 asymptomatic individuals of the Framingham Heart Study (103). Similar to what has been previously reported with EAT, PAT was positively correlated with BMI and waist circumference, but the correlation with VAT accumulation was of slightly higher magnitude. PAT was associated with a whole set of cardiometabolic risk factors related to the metabolic syndrome independently of BMI and waist circumference, but not independently of VAT. Although the relationship of PAT with coronary artery calcium was independent of VAT, VAT was closely associated with the prevalence of type 2 diabetes, impaired fasting glucose, hypertension, and metabolic syndrome at any level of PAT. Another recent report from Framingham that included individuals with CVD (n = 1,267) suggested that independently of BMI and waist circumference both PAT and VAT were associated with CVD, with VAT again showing correlations with CVD of higher magnitude than PAT (104). Similar findings were observed when magnetic resonance imaging (MRI) measures of left ventricular mass and left ventricular end-diastolic volume were considered as outcomes (105). Finally, using proton magnetic resonance spectroscopy, McGavock et al. (106) have shown that the lipid content of cardiomyocytes was another hall-mark feature of type 2 diabetes that was closely associated with both serum and hepatic triglyceride levels in humans. This study confirmed findings that have been previously reported in the Zucker diabetic fatty rat which revealed several cardiac abnormalities such as eccentric left ventricular remodeling, increased left ventricular pressure, and decreased systolic efficiency, most of which were reversed with thiazolidinedione treatment (107,108). In humans, Kankaanpaa et al. have also reported that myocardial triglyceride content is associated with plasma FFA levels, overall ectopic fat deposition, and left ventricular hypertrophy (109). On the impact of exercise training on EAT and VAT, Kim et al. (110) have shown that this type of intervention promoted a profound loss of EAT (twice as high as the relative percentage loss of either waist circumference orBMI) and that this loss in EAT was positively associated with loss of VAT. Moreover, as recently pointed out by Sacks (111), it is tempting to suggest that EAT promotes the secretion of a VAT-like adipokine pattern. However, this has not been proven in humans. The literature on the topic is somewhat scarce, and more studies on this topic are needed, especially in apparently healthy non-obese humans. In this context, a study showed that visceral but not epicardial fat is associated with blood pressure indices and severity of insulin resistance in a small sample of non-diabetic men (112).

Similar to VAT, fat depots located in the heart may contribute to the low-grade inflammatory state that characterizes individuals with ectopic fat deposition/metabolic syndrome (113). For instance, patients with documented coronary artery disease exhibit higher levels of IL-6 gene expression while showing lower levels of adiponectin expression (114). Similar findings have been observed in patients with hypertension (115). Interestingly, a recent study conducted in a population with metabolic syndrome and/or type 2 diabetes has documented the benefits of pioglitazone in reducing the genetic expression of a number of pro-inflammatory mediators such as IL-1β, IL-1Ra, and IL-10 (116). Taken together, these observations suggest that ectopic fat deposition in various regions of the heart is detrimental to energy homeostasis. EAT, PAT, and intramyocellular lipids (IMCL) may be important causes of local insulin resistance and promote the secretion of several adipokines that may strongly impact the heart's ability to remove plasma lipids and use them as an energy source.

Intramyocellular lipids

Muscular infiltration of lipids is not limited to the heart muscle as virtually all skeletal muscles are subject to lipid infiltration, which may impair energy homeostasis and create a local state of muscular insulin resistance. Two fat depots are found in skeletal muscle: intramyocellular ‘intramuscular’ fat, which is fat infiltration within myocytes, and extramyocellular ‘intermuscular’ fat, which is the small adipose depot interspersed between muscle fibers. Although lipid droplets inside the skeletal muscle may be an important source of FFAs to be oxidized by surrounding cells, an elevated IMCL content assessed by proton nuclear magnetic resonance (NMR) spectroscopy has been shown to be a close correlate of insulin resistance in several studies (117,118). The ability to distinguish between the two depots may be important since studies have suggested that extramyocellular fat is also associated with higher fasting glucose (119), impaired glucose tolerance, insulin resistance, type 2 diabetes (120), and metabolic syndrome (121), independent of overall obesity. Interestingly, a recent longitudinal study has shown that skeletal muscle fat infiltration increases with aging and decreases with exercisetraining. Notably, such changes are mostly independent of changes in body-weight(122). However, for the moment, there are few data available on a potential role for IMCL as an important risk factor for CVD or type 2 diabetes. As IMCL is usually correlated with intrahepatic fat accumulation or VAT, the deteriorated cardiometabolic risk profile has been proposed to be most likely attributable to these ectopic deposition sites rather than to IMCL per se (123). The benefits of PPAR-γ activation have also been reported in a study by Bajaj et al. (124) who have shown that following pioglitazone treatment in a population of patients with type 2 diabetes, intramyocellular fat content (measured by proton NMR) decreased by approximately 40%.

Pancreatic fat

In the natural development of type 2 diabetes, insulin secretion generally increases as insulin resistance develops until the pancreas can no longer secrete enough insulin to maintain glycemia to a normal level. However, ectopic lipid accumulation in the pancreas may represent a key feature of obesity-driven type 2 diabetes and could also explain to a certain extent such impairment in insulin secretion (125). A number of studies have recently investigated the potential relationship between pancreas lipid accumulation and β-cell dysfunction. Animal studies have shown that increased pancreatic lipid accumulation was associated with deterioration of both β-cell mass and function (126,127). In humans, lipid-induced β-cell damage could be present for more than a decade before the onset of type 2 diabetes. Pancreatic lipid content increases with deterioration of glucometabolic state, as shown by van der Zijl et al. (128). Another study in humans has also shown that increased pancreatic fat directly associates with impaired insulin secretion in subjects with impaired fasting glucose (129). Interestingly, it has been shown that mice with specific inactivation of ATP binding-cassette A-1 (ABCA1), a cellular cholesterol transporter, in the β-cell have impaired glucose tolerance and that,ex vivo, the islets of ABCA1^-/- mice showed impaired insulin secretion (130). In humans, carriers of loss-of-function mutations in ABCA1 have impaired pancreatic β-cell function, which suggests that the cholesterol metabolism may play a significant role in the etiology of insulin resistance (131). In humans, the concept of lipotoxicity as a causal factor for β-cell dysfunction remains unclear and warrants further investigation (132).

In comparison with other fat sites, fat infiltrated in the heart, skeletal muscle, and the pancreas accounts for a very small proportion of total body fat (less than 5%). However, the local toxic effects may play a very important pathophysiological role in the various systems that regulate energy homeostasis. Consequently, removal or utilization of lipids in these ectopic sites may become an important target of therapy to prevent the onset of chronic diseases such as type 2 diabetes, CVD, and hypertension (Figure 1). Experimental and clinical studies are clearly needed to test this hypothesis. Additionally, although the relationships between fat accumulation in various tissues are often correlated to oneanother, studies have suggested that this may not always be the case (106). One of the biggest challenges of the field will be to validate whether or not ectopic fat in the above-discussed depots will predict CVD and/or type 2 diabetes independent of each other.

Figure 1. Schematic representation of abnormal visceral adipose tissue development leading to the storage of lipids at undesired sites. This phenomenon has been described as ‘ectopic fat’ deposition. According to this model, when energy intake is greater than energy expenditure, a dysfunctional subcutaneous adipose tissue may be unable to store appropriately the dietary energy excess, eventually leading to lipid overflow and to visceral fat deposition. This excess of visceral adiposity is associated with an altered free fatty acid metabolism with the release of adipokines as well as with adipocyte hypertrophy and increased fat storage in non-adipose organs, notably liver and skeletal muscle, and in other organs such as the pancreas and the heart. Excess adipocyte hypertrophy leads to ectopic fat storage and metabolic abnormalities such as insulin resistance, atherogenic dyslipidemia, hypertension, systemic inflammation, and eventually increases the risk of CVD.

Assessing body fat distribution with imaging techniques

Several imaging techniques such as CT and MRI have been proven to be useful for the determination of total and regional body composition. At the time being, these methods are the most accurate tools available for directly quantifying body composition at a tissue level. Both methods provide cross- sectional images that can be used to determine SAT and VAT accumulation in vivo. CT uses ionizing radiation and differences in tissue X-ray attenuation to produce cross-sectional images of the body, which can be compared to determine the surface of various fat depots (12,133). Although excess VAT accumulation is increasingly recognized as a risk factor for type 2 diabetes and CVD, there is currently no consensus with regard to optimal cut-offs for identifying individuals with the high-risk obesity phenotype. Currently, suggested cut-offs at the L4–L5 level range from 100 to 130 cm². Nevertheless, regardless of optimal cut-off values, an increase in VAT is a hall-mark feature of increased cardiometabolic risk, even among normal-weight individuals. However, as CT is expensive and involves radiation exposure, it is currently not advised to measure routinely the VAT using CT or MRI in clinical practice. Health professionals are therefore advised to focus on measuring waist circumference on a regular basis as waist circumference is, for the time being, the best surrogate measure currently available. CT can also be used to assess lipid infiltration in other peripheral tissues such as muscle and the liver (134). Since fat has a lower density than both water and protein, visceral, muscle, or liver fat infiltration can be reflected and quantified by a lower liver density and thus lower attenuation values.

Glycemic control and CVD risk: are we aiming at the right target?

There is now considerable evidence that type 2 diabetes is a significant risk factor for CVD. Although this is not a consistent finding, some prospective studies have reported that individuals with type 2 diabetes without CVD may carry the same risk of incident CVD as individuals without diabetes that have previously suffered a myocardial infarction (135,136). Clinicians are aware of the deleterious vascular consequences of a poorly controlled hyperglycemic state. For instance, a poor glycemic control reflected by high plasma glycated hemoglobin levels is associated with microcirculatory complications which may lead to retinopathies, nephropathies, and neuropathic complications (137,138). These microcirculatory damages explain why individuals with poor glycemic control are at increased risk for blindness and end-stage renal disease, which may lead to dialysis, as well as amputations (139).

On the basis of clinical trials that have documented the protective effects of a better glycemic control on the microcirculation, physicians are required to use a broad pharmacological arsenal and, if needed, exogenous insulin to protect patients against these complications (140). However, the literature providing evidence for a strong protective effect of better glycemic control on atherosclerotic macrovascular disease is limited. For instance, although the pharmacological management of concomitant risk factors such as hypertension and LDL-cholesterol levels has clearly produced cardiovascular benefits in patients with type 2 diabetes, three recently reported trials have failed to report clear cardiovascular benefits of a better glycemic control achieved by intensive pharmacotherapy (141–143). Although results coming from these large clinical studies may be disappointing at first sight, they have led the scientific and medical community to seek additional risk factors for CVD and other therapeutic targets in patients with diabetes. In this regard, we and other groups around the world have shown that the high triglyceride–low HDL cholesterol atherogenic dyslipidemia, which is part of a broader constellation of risk factors including visceral obesity and inflammation, is predictive of a further increase in CVD risk in patients with type 2 diabetes (144–146). Therefore, even if the population of patients with type 2 diabetes is, as a whole, characterized by an increased CVD risk, a better ‘phenotyping’ of patients with type 2 diabetes would require going beyond the assessment of glycemic control and of traditional CVD risk factors. In this regard, studies have shown that approximately 15% of patients with type 2 diabetes are neither viscerally obese nor dyslipidemic and are therefore at much lower risk of CVD than the majority of patients (85%) with type 2 diabetes who are likely to carry a certain amount of VAT and/or ectopic fat (147). Furthermore, within the latter group, we have proposed that the greater the amount of VAT/ectopic fat, the greater will be the risk of CVD among patients with type 2 diabetes (144). It is increasingly recognized that type 2 diabetes is a very heterogeneous condition and that those patients with type 2 diabetes that are at higher CVD risk are likely to be those with deleterious body fat distribution patterns. These patients nowadays unfortunately represent the vast majority of patients with type 2 diabetes.

In order to decrease the odds of apparently healthy individuals with visceral obesity/ectopic fat converting to type 2 diabetes or having a vascular event, increasing daily energy expenditure though regular physical activity/exercise is warranted in most if not all individuals. In this context, the results of the Diabetes Prevention Program have shown that increasing physical activity levels delays the onset of both type 2 diabetes (148) and metabolic syndrome (149). Interestingly, the recently published results of the LOOK-AHEAD trial (150) may suggest that increasing physical activity levels improves cardiovascular risk factors in obese patients with type 2 diabetes. On top of life-style therapy, careful management of parameters of the lipoprotein-lipid profile is advised in this high-risk population, and both statin and fibrate therapy have been proven effective in patients with metabolic syndrome or those with the hightriglyceride–low HDL cholesterol atherogenic dyslipidemia (151–154).

Conclusions

Although mortality from CVD appears to have recently decreased in several developed countries, there is evidence that the mosaic of modifiable risk factors for CVD has evolved with an increased prevalence of type 2 diabetes that is likely to be attributable to the increased prevalence of individuals characterized by high amounts of visceral fat (155,156). As the worldwide prevalence of patients characterized by a deteriorated cardiometabolic risk profile is steadily increasing (157,158), there is an urgent need to develop new therapeutic approaches where the loss of visceral/ectopic fat would become an additional therapeutic target which would be as important if not more as achieving optimal glycemic control (Figure 2). With the arsenal of non-invasive imaging techniques to assess not only body fat distribution, there is a remarkable opportunity to decipher better the key drivers of cardiometabolic risk in a rapidly expanding group of individuals with poor life-style habits and with an excess of visceral/ectopic fat.

Figure 2. Schematic representation of the assessment of global cardiometabolic risk using traditional risk factors and emerging cardiovascular disease risk markers.

Acknowledgements

The work of the authors has been supported by research grants from the Canadian Institutes of Health Research, the Heart and Stroke Foundation, and by the Foundation of the Quebec Heart Institute. Dr Larose is a researchscholar of the Fonds de la Recherche en Santé du Québec (FRSQ). Dr Després is the Scientific Director of the International Chair on Cardiometabolic Risk, which is based at the Université Laval. Benoit J. Arsenault is supported by a post-doctoral fellowhip from the Fonds de la Recherche en Santé du Québec and the Fondation de l’Institut universitaire de cardiologie et de pneumologie de Québec.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.