PEOPLE WITH TYPE 2 diabetes are disproportionately

Transcription

1 X/04/$20.00/0 Endocrine Reviews 25(1): Printed in U.S.A. Copyright 2004 by The Endocrine Society doi: /er Nontraditional Risk Factors for Cardiovascular Disease in Diabetes V. FONSECA, C. DESOUZA, S. ASNANI, AND I. JIALAL Section of Endocrinology (V.F., C.D., S.A.), Department of Medicine, Tulane University Medical Center, New Orleans, Louisiana 70112; and Laboratory for Atherosclerosis and Metabolic Research (I.J.), University of California Davis, Sacramento, California People with type 2 diabetes are disproportionately affected by cardiovascular disease (CVD), compared with those without diabetes. Traditional risk factors do not fully explain this excess risk, and other nontraditional risk factors may be important. This review will highlight the importance of nontraditional risk factors for CVD in the setting of type 2 diabetes I. Introduction A. Potential role of nontraditional risk factors for cardiovascular disease (CVD) in type 2 diabetes B. Natural history of vascular abnormalities in the pathogenesis of CVD in diabetes II. Hyperinsulinemia/IR A. Epidemiology and clinical observations B. Proposed mechanisms linking IR with CVD in diabetes III. Endothelial Dysfunction A. Epidemiology and clinical observations B. Proposed mechanisms linking endothelial dysfunction with CVD in diabetes IV. Impaired Fibrinolysis and Prothrombotic State A. Epidemiology and clinical observations B. Proposed mechanisms linking impaired fibrinolysis and thrombosis with CVD in diabetes V. Inflammation and CVD in Diabetes A. Epidemiology and clinical observations B. Proposed mechanisms linking inflammation with CVD in diabetes VI. Microalbuminuria Abbreviations: ABI, Ankle-brachial pressure index; ACE, angiotensin-converting enzyme; ADMA, asymmetric dimethylarginine; AT, -tocopherol; BH4, tetrahydrobiopterin; CHD, coronary heart disease; CRP, C-reactive protein; CVD, cardiovascular disease; EBCT, electron beam-computed tomography; enos, endothelial nitric oxide synthase; FFA, free fatty acid; FMD, flow-mediated dilation; HDL, high-density lipoprotein; hs, high sensitive; ICAM, intercellular adhesion molecule; IGT, impaired glucose tolerance; IMT, intima-media thickness; IR, insulin resistance; IRS, insulin resistance syndrome; LDL, low-density lipoprotein; NCEP-ATP, National Cholesterol Education Program Adult Treatment Panel; NO, nitric oxide; NOS, NO synthase; Ox-LDL, oxidized LDL; PAI-1, plasminogen activator inhibitor type 1; SI, insulin sensitivity; thcy, total homocysteine; TZD, thiazolidinedione; VCAM, vascular cell adhesion molecule; VLDL, very LDL; vwf, von Willebrand factor. Endocrine Reviews is published bimonthly by The Endocrine Society ( the foremost professional society serving the endocrine community. 153 and discuss their role in the pathogenesis of the excess CVD morbidity and mortality in these patients. We will also discuss the impact of various therapies used in patients with diabetes on nontraditional risk factors. (Endocrine Reviews 25: , 2004) A. Epidemiology and strength of the association of microalbuminuria with CVD in diabetes B. Proposed mechanisms linking microalbuminuria with CVD in diabetes VII. Hyperhomocysteinemia A. Epidemiology and strength of the association of homocysteine with CVD in diabetes B. Proposed mechanisms linking homocysteine with CVD in diabetes VIII. Vascular Wall Abnormalities Carotid IMT and Arterial Stiffness A. Epidemiology and clinical observations B. Proposed mechanisms linking vascular wall abnormalities with CVD in diabetes IX. Postprandial Hyperglycemia A. Epidemiology and clinical observations relating postprandial hyperglycemia to CVD in diabetes B. Proposed mechanisms linking postprandial hyperglycemia with CVD in diabetes X. Impact of Various Treatments on Nontraditional Risk Factors for CVD in Diabetes A. Lifestyle change B. Glycemic control C. Insulin sensitizers D. Antiplatelet therapy E. Blocking angiotensin formation and action F. Effect of lipid-lowering therapy on nontraditional risk factors G. Vitamins C and E H. Homocysteine-lowering treatment XI. Summary and Clinical Implications I. Introduction A. Potential role of nontraditional risk factors for cardiovascular disease (CVD) in type 2 diabetes PEOPLE WITH TYPE 2 diabetes are disproportionately affected by CVD, compared with those without diabetes. Furthermore, diabetic patients have not benefited from

2 154 Endocrine Reviews, February 2004, 25(1): Fonseca et al. Cardiovascular Risk Factors in Diabetes the advances in the management of coronary heart disease (CHD) and/or its risk factors that have resulted in a decrease in mortality for CHD patients without diabetes (1). Some of this excess risk relates to an increased prevalence of established risk factors, such as obesity, dyslipidemia, and hypertension, in persons with diabetes. The relative importance of these risk factors has been extensively reviewed in the literature and was highlighted in the United Kingdom Prospective Diabetes Study (UKPDS) (2). Nevertheless, these traditional risk factors do not fully explain the excess risk for CHD associated with diabetes (3). Therefore, other nontraditional risk factors (Table 1) may be important in people with diabetes (4, 5). Several studies have demonstrated the importance of these factors in the pathogenesis of CVD in diabetes. However, very few of them have demonstrated prospectively the associations of nontraditional risk factors in diabetes, independent of traditional risk factors (5). Several therapeutic strategies that are currently used in the management of diabetes can improve these nontraditional risk factors. For example, lipid management with statins reduces markers of inflammation in plasma (6). Similarly, insulin sensitizers have a variety of effects on many of these risk factors (7). The purpose of this review is to highlight the importance of nontraditional risk factors for CVD in the setting of type 2 diabetes and discuss their role in the pathogenesis of the excess CVD morbidity and mortality in these patients. Finally, we will discuss the impact of therapy of diabetes on nontraditional risk factors. It is important to recognize that these risk factors do not function in isolation. In fact, they frequently cluster in individual patients and possibly interact with each other, although such interactions are hard to quantify. Figure 1 represents a hypothetical model of interaction of various risk factors. Although some of these risk factors cluster, others appear to be independent of each other. Furthermore, studies on these risk factors are usually either epidemiological descriptions of associations with CVD or experimental studies focusing on a single risk factor. We recognize that CVD is a complex multifactorial disease and that many of these processes are functioning simultaneously (8). Nevertheless, in TABLE 1. Risk factors for CVD in diabetes Traditional Hypertension Dyslipidemia Family history of premature CVD Cigarette smoking Nontraditional IR Endothelial dysfunction 2 Vascular reactivity, 2 NO 2 ADMA Impaired fibrinolysis 1 PAI-1 Inflammation 1 hs-crp, 1 WBC 1 Adhesion molecules, 1 MMP-9 Microalbuminuria Hyperhomocysteinemia Postprandial abnormalities Vascular wall abnormalities 1 IMT, calcification, 2 compliance 1, Increased; 2, decreased; WBC, white blood cells; MMP-9, matrix metalloproteinase 9. the interest of simplicity, we will consider each risk factor separately. These risk factors provide insights into the pathogenesis of CHD in diabetes. Many of the risk factors identified may reflect underlying inflammation, which may be either secondary to atherosclerosis itself, or may represent a nonspecific inflammatory response to infections, which are likely to be much more common in people with uncontrolled diabetes. Even if the origins of these risk factors are nonspecific, they may be important, because they have the potential to enhance the coagulation and thrombotic process, thus perpetuating the vascular disease or triggering vascular events. Many of these risk factors may be common antecedents for both diabetes and CHD, supporting the hypothesis that both disorders arise separately from common soil (9). Possible common antecedents include insulin resistance (IR) and inflammation, which have been shown to be independent risk factors for CHD and are clearly associated with type 2 diabetes. Consistent with the common soil hypothesis, some of these nontraditional risk factors have also predicted incident diabetes in the same studies, suggesting that they are altered in the prediabetic state, during which both IR and inflammation are often present. Inflammation, endothelial dysfunction, and abnormalities of coagulation are all associated with IR and may thereby become common antecedents of both diabetes and CHD (9). B. Natural history of vascular abnormalities in the pathogenesis of CVD in diabetes The current concepts with regard to atherosclerosis suggest that the earliest event in atherogenesis is endothelial cell dysfunction manifesting as deficiencies of nitric oxide (NO) and prostacyclin. This can be induced by various noxious insults including dyslipidemia, diabetes, hypertension, smoking, etc. Recent data suggest that the prediabetic state may be associated with endothelial dysfunction possibly due to IR. The next event in atherogenesis is the binding of mononuclear cells, such as monocytes and T lymphocytes, to the endothelium; this binding is mediated by adhesion molecules present on the endothelial surface, such as vascular cell adhesion molecule (VCAM), intercellular adhesion molecule (ICAM), and E-selectin. Once the monocyte migrates into the subendothelial space, it matures into a resident macrophage, takes up lipid largely through certain scavenger receptors such as SR-A and CD-36, and becomes a foam cell. In the later stages of atherogenesis, smooth muscle cells migrate to the surface and form the fibrous cap of the lesion. Finally, lipidladen macrophages release matrix metalloproteinases causing plaque rupture and acute coronary syndromes such as myocardial infarction and unstable angina. Oxidative stress plays a crucial role in atherogenesis, especially in diabetes (10, 11). Several lines of evidence support a proatherogenic role for oxidized low-density lipoprotein (Ox-LDL) and its in vivo existence (12, 13). Ox-LDL is not recognized by the LDL receptor but by the scavenger receptor pathway on macrophages, which results in unregulated cholesterol accumulation, leading to foam cell formation. Factors that may promote increased oxidative stress in diabetes include antioxidant deficiencies, increased production

3 Fonseca et al. Cardiovascular Risk Factors in Diabetes Endocrine Reviews, February 2004, 25(1): FIG. 1. Interaction of nontraditional risk factors in diabetes. of reactive oxygen species, and the process of glycation and glycooxidation (13). Direct evidence of increased oxidative stress and lipid peroxidation in diabetes has been reported and recently reviewed (11). Clinical markers include F2-isoprostanes, which are prostaglandin-like compounds formed in vivo from free radical-catalyzed peroxidation of arachidonic acid and have emerged as novel and direct measures of oxidative stress. F2-isoprostane levels are increased in both the urine and plasma of patients with type 2 diabetes (14 16). Also, patients with type 2 diabetes have elevated plasma levels of nitrotyrosine, another marker of protein oxidation (14, 15), as well as evidence of oxidative damage to DNA (17). In summary, CVD is increased in diabetes due to a complex interplay of many traditional and nontraditional risk factors. The latter have recently been recognized to have a significant role in the initiation and the progression of atherosclerosis, over its long natural history from endothelial function to clinical events. II. Hyperinsulinemia/IR A. Epidemiology and clinical observations The insulin resistance syndrome (IRS), also known as the metabolic syndrome, is a cluster of cardiovascular risk factors frequently, but not always, associated with obesity. Reaven (18, 19) first drew attention to the association of IR and obesity, type 2 diabetes, high plasma triglycerides, and low plasma high-density lipoprotein (HDL) cholesterol. Since its original description, there has been much experimental, clinical, and epidemiological data to support the association of this syndrome with CVD (20). Additionally, other cardiovascular risk factors have been frequently included in the description of the syndrome. These include inflammation, abnormal fibrinolysis, and endothelial dysfunction (4). It is unclear to what extent the components of this syndrome develop independently of each other or spring from common soil genetic abnormalities (9). Due to the increase in the prevalence of obesity and the frequent coexistence of these abnormalities in obese subjects, the syndrome has become a major clinical and public health problem (21). Unfortunately, it is difficult to measure IR in epidemiological studies, particularly in patients with advanced type 2 diabetes. Research studies have used complex experimental techniques to quantify insulin sensitivity (SI)/resistance. Epidemiological studies utilize hyperinsulinemia to define IR. Studies use either fasting plasma insulin alone or formulas based on plasma insulin and glucose (such as the homeostasis model assessment, commonly known as HOMA). Because plasma insulin concentrations are a reflection of both ambient glucose and pancreatic -cell function (which decreases even before the onset of type 2 diabetes), it is a poor marker of IR. Furthermore, lack of standardization of the insulin assay makes interpretation of plasma insulin concentration difficult. More, recently the World Health Organization (WHO) and the National Cholesterol Education Program Adult Treatment Panel III (NCEP-ATP III) have attempted to define the syndrome for clinicians (NCEP-ATP III internet site Subjects identified using these clinical definitions have been shown to be at increased risk of CVD. Prospective studies suggest that hyperinsulinemia may be an important risk factor for ischemic heart disease. The Quebec Heart Study studied men who were 45 to 76 yr of age and who did not have ischemic heart disease (22). A first ischemic event occurred in 114 men who were then matched for age, body mass index, smoking habits, and alcohol consumption with a control selected from among the 1989 men who remained free of ischemic heart disease during follow-up. Fasting insulin concentrations at baseline were 18% higher in the study patients than in the controls. High fasting insulin concentrations were an independent predictor of ischemic heart disease in these men after adjustment for systolic blood pressure, family history of ischemic heart disease, plasma triglyceride, apolipoprotein B, LDL cholesterol, and HDL cholesterol concentrations. Similarly, hyperinsulinemia was associated with increased all-cause and cardiovascular mor-

4 156 Endocrine Reviews, February 2004, 25(1): Fonseca et al. Cardiovascular Risk Factors in Diabetes tality in Helsinki policemen independent of other risk factors (23). Because correlations of insulin with other risk factors make interpretation difficult, factor analysis to study the clustering of risk factors in the baseline data of the Helsinki Policemen Study was carried out. Factor analysis including only risk factor variables proposed to be central components of IRS predicted the risk of CHD and stroke independently of other risk factors (24). Other population studies have supported these studies (25 27). Nevertheless, it is important to recognize that the relationship between IR and plasma insulin may not be linear (28), and some studies have been negative (29). A recent metaanalysis, however, cautioned that the association was somewhat weak, although statistically significant (25). The metaanalysis resulted in an estimated summary relative risk of 1.18 for differences in insulin level, equivalent to the difference between the 75th and 25th percentiles of the general population. Ethnic background and type of insulin assay modified the relationship between insulin and CVD. Overall, these studies confirm that components of the syndrome are present for several years before the onset of type 2 diabetes and that the clock for coronary heart disease starts ticking before the onset of clinical diabetes (27). B. Proposed mechanisms linking IR with CVD in diabetes The exact mechanism by which IR causes CVD is not known. However, IR is associated with several other cardiovascular risk factors, some of which are discussed below. Insulin-resistant type 2 diabetic subjects have more atherogenic cardiovascular risk factor profiles than insulin-sensitive type 2 diabetic subjects, and this is only partially related to increased obesity and an adverse body fat distribution (30). 1. Obesity. Obesity is frequently associated with several of the components of the IRS and may be critical for the development of the syndrome. Several mechanisms have been proposed for the link between obesity and the IRS (31). Cardiovascular morbidity and mortality are increased in obese individuals independently of other risk factors. IR is very common in obese individuals. However, some nonobese individuals demonstrate hyperinsulinemia and the other features of the IRS (32). Thus, obesity may not be essential for the expression of the syndrome, but the presence of obesity or weight gain may accentuate the pathophysiological changes associated with the syndrome. Body fat distribution rather than body mass may actually be a better predictor of IR and cardiovascular risk (33). IR, type 2 diabetes, and hypertension are more closely associated with a central distribution of adiposity than with general increases in fat mass. Waist circumference serves as a clinical surrogate of intraabdominal fat and correlates with insulin levels and IR. Adipose tissue is now recognized to be a significant endocrine organ secreting a variety of hormones and cytokines. Data suggest that some of these cytokines arising from adipose tissue may be partly responsible for the metabolic, hemodynamic, and hemostatic abnormalities associated with IR. Studies show a close relationship between obesity and circulating C-reactive protein (CRP), TNF, and IL-6, and some of these cytokines are predictors of CVD. Plasma CRP is elevated in obese subjects who have other features of the IRS (34). Thus, inflammation originating from excess adipose tissue cytokine production may contribute not only to the development of the IRS but also to the associated CVD. Increased expression of TNF in adipose tissue has been reported in obese subjects. TNF inhibits the action of lipoprotein lipase and stimulates lipolysis. TNF promotes monocyte adhesion to the endothelium and inhibits endothelial nitric oxide synthase (enos). TNF also impairs the function of the insulin-signaling pathway by effects on phosphorylation of both the insulin receptor and insulin receptor substrate-1. IL-6 may also induce endothelial expression of cytokines, thereby contributing to endothelial dysfunction. 2. Dyslipidemia. Hyperlipidemia is well established as a risk factor in diabetics to the same extent as in nondiabetics. However, certain qualitative abnormalities in the lipoprotein pattern associated with IR appear to convey excess risk and could be classified as nontraditional risk factors. One of the characteristic relationships between IR and a cardiovascular risk factor is with diabetic dyslipidemia (35). The hallmark of the syndrome is hypertriglyceridemia and low plasma HDL cholesterol concentration. Plasma LDL cholesterol concentrations in insulin-resistant subjects are no different from those in insulin-sensitive subjects. However, there are qualitative changes in LDL cholesterol resulting in pattern B distribution of LDL particles, which consists of smaller LDL particles that are more susceptible to oxidation and thus potentially more atherogenic (36). Small dense LDL particles permeate the arterial wall faster and bind more avidly to proteoglycans than larger LDL particles. IR at the level of adipose tissue may result in increased activity of hormone-sensitive lipase and therefore increased breakdown of stored triglycerides. Free fatty acids (FFAs) released from adipocytes, particularly intraabdominal adipocytes, can be transported to the liver where they stimulate synthesis of triglycerides and assembly and secretion of very LDL (VLDL). Increased plasma VLDL triglycerides exchange with cholesterol esters from HDL, resulting in a lower plasma HDL cholesterol. On the other hand, an increase in circulating FFA has been proposed as having an etiological role in the development of IR (37). The effect of treatment of IR on dyslipidemia is discussed in Section II.B Hypertension. Although it is well established that essential hypertension is frequently associated with IR, the impact of this abnormality on blood pressure homeostasis is still a matter of debate. Fasting plasma insulin is frequently higher in hypertensive subjects, and glucose disposal during an euglycemic clamp is decreased. The association between hypertension and IR is more convincing in obese subjects. Significant decreases in blood pressure have been observed in obese subjects who lose modest amounts of weight, correlating closely with the decline in fasting plasma insulin concentrations. Plasma insulin concentrations are higher and insulin-mediated total-body glucose disposal is reduced in young, normal weight individuals with essential hypertension (20). The impairment in insulin-mediated glucose dis-

5 Fonseca et al. Cardiovascular Risk Factors in Diabetes Endocrine Reviews, February 2004, 25(1): posal was closely related to the increase in blood pressure. Multiple potential mechanisms by which IR may cause hypertension have been proposed (38). These include resistance to insulin-mediated vasodilation, impaired endothelial function, sympathetic nervous system overactivity, sodium retention, increased vascular sensitivity to the vasoconstrictor effect of pressor amines, and enhanced growth factor activity leading to proliferation of smooth muscle walls. However, some studies do not support the association of metabolic IR with essential hypertension. Clearly, hypertension is itself a complex disorder with many etiologies, and not all subjects with essential hypertension are insulin resistant. 4. Abnormal insulin signaling, hyperinsulinemia, and the vasculature. As outlined above, IR with resultant hyperinsulinemia is an independent risk factor for CVD. It is important to recognize that although hyperinsulinemia is frequently used as a surrogate marker for underlying IR, it is still controversial whether insulin itself is the culprit, and the specific role of insulin in the pathogenesis of atherosclerosis remains unclear. In fact, insulin has vasodilatory and antiinflammatory properties, which should protect against atherosclerosis. Several mechanistic hypotheses have been proposed to explain this controversy (20, 39). First, insulin is a growth factor that stimulates vascular cell growth and synthesis of matrix proteins. Second, the insulin signaling pathway thought to be responsible for abnormalities in glucose metabolism is also involved in NO production. Thus, the abnormal intracellular signaling that causes hyperglycemia may also be responsible for vascular disease due to loss of insulin s antiatherogenic properties, whereas hyperinsulinemia continues to stimulate growth-promoting enzymes such as MAPK (39). Although some controversy remains, this hypothesis has been supported by many studies. In addition, imbalances in insulin homeostasis are associated with abnormalities in expression and action of various peptides, growth factors, and cytokines. These include angiotensin II, endothelin-1, and IGF-I (39). Although the exact role of peroxisome proliferatoractivated receptors in the pathogenesis of this syndrome is unclear, several studies support the concept that they may have a role in the development of not only IR but also atherosclerosis (40). For example, these receptors are present in vascular tissue, heterozygous mutations in the ligand-binding domain of peroxisomal proliferator-activated receptor- are occasionally associated with IR, and agonists of these receptors have a significant impact on the syndrome. In summary, IR is a key abnormality linking type 2 diabetes and CVD. By its clinical definition, it is associated closely with traditional cardiovascular risk factors. As discussed in Section III, it is also significantly related to nontraditional risk factors as well. III. Endothelial Dysfunction The importance of the endothelium in maintaining vascular health has been widely recognized. The endothelium is a critical determinant of vascular tone, reactivity, inflammation, vascular remodeling, maintenance of vascular patency, and blood fluidity. The importance of endothelial dysfunction in the pathogenesis of CVD in diabetes has been recognized only recently (41), and, in fact, it may be a prognostic/risk marker (42). Many of the functions of the endothelium are maintained through paracrine and endocrine regulatory substances secreted from endothelial cells, which may often have opposing actions. For example, endothelial cells secrete NO, the most potent known vasodilator. Endothelial cells also secrete other important vasodilators such as prostacyclin. The vasodilatory actions are opposed by secretion of potent vasoconstrictors such as endothelin 1. Similarly, these and other endothelial products are involved in maintaining the balance between smooth muscle cell growth, promotion and inhibition, thrombosis and fibrinolysis, inflammation, and cell adhesion. Assessment of endothelial function can be broadly divided into biochemical and functional. Biochemical parameters of endothelial dysfunction have frequently been described as risk factors for CVD and include plasma von Willebrand factor (vwf), thrombomodulin, and several adhesion molecules such as VCAM, ICAM, E-Selectin, and P-Selectin. Functional assessment is dependent on the ability of blood vessels to dilate in response to a number of varied stimuli, such as shear stress and acetylcholine infusion. These stimuli result in the release of NO from the endothelium and therefore measure endothelium-dependent vasodilation. Details of the various methods used to assess endothelial function are beyond the scope of this review. The ability of blood vessels to dilate in response to stimuli, including ischemia, is called vascular reactivity or flowmediated dilation (FMD). Brachial artery vascular reactivity is a noninvasive method of assessing arterial endothelial function in vivo. Because endothelial injury is an early event in atherogenesis, it has been suggested that abnormal FMD may precede the development of structural changes in the vessel wall. Abnormal FMD has been shown in several insulin-resistant states and is present in relatives of patients with type 2 diabetes who have normal glucose tolerance. It has even been proposed that endothelial dysfunction may be a precursor of the IRS (43). Table 2 lists various endothelial abnormalities associated with IR. A. Epidemiology and clinical observations Endothelial dysfunction is now recognized as being an early abnormality in the natural history of CVD and may be a good predictor of cardiovascular events (44 46). Its importance in diabetes is now well recognized (41). NOmediated vasodilation is impaired in non-insulin-dependent diabetes mellitus (47). This is possibly an intrinsic abnormality related to IR, although hyperglycemia may exacerbate endothelial dysfunction (48). Impaired endothelial function was first associated with CVD when Ludmer et al. (49) injected acetylcholine into coronary arteries and demonstrated paradoxical vasoconstriction, instead of the expected vasodilation. Schachinger et al. (44) assessed coronary endothelial vasomotor function in response to acetylcholine and nitroglycerin infusion in 147 patients undergoing cardiac catheterization. During a 7.7-yr follow-up, 16 patients had a cardiovascular event, and en-

6 158 Endocrine Reviews, February 2004, 25(1): Fonseca et al. Cardiovascular Risk Factors in Diabetes TABLE 2. Alterations in the vascular endothelium associated with diabetes mellitus and IR Abnormality Significance 2 Release of and responsiveness to NO Impaired endothelial function and reactivity 1 Expression, synthesis, and plasma levels of endothelin-1 Vasoconstriction and hypertension 2 Prostacyclin release Impaired vasodilation 1 Adhesion-molecule expression Increased monocyte adhesion to vessel wall 1 Adhesion of platelets and monocytes Foam cell formation, thrombosis, and inflammation 1 Procoagulant activity Thrombosis 1 Advanced glycosylated end products Increased stiffness of arterial wall Impaired fibrinolytic activity Decreased clot breakdown 1, Increased; 2, decreased. dothelial-dependent and -independent vasodilation was significantly worse in these 16 patients (44, 45). Suwaidi et al. (49a) tested endothelial function in 157 patients with mild coronary atherosclerosis. During a 28-month follow-up, six patients had cardiovascular events, and all six had prior evidence of severe endothelial dysfunction. A recent large study has demonstrated that epicardial and microvascular coronary endothelial dysfunction independently predict acute cardiovascular events in patients with and without coronary artery disease, providing both functional and prognostic information that complements angiographic and risk factor assessment (50). Finally, impaired brachial artery reactivity independently predicts postoperative cardiac events such as myocardial infarction, unstable angina, stroke, and cardiac death in patients undergoing vascular surgery (51). It has become clear that among its many actions insulin is also a vasoactive hormone (52 56). Its effect to cause endothelial-no-dependent vasodilation is physiological and dose dependent (57). Insulin has been shown to induce expression of the enzyme NO synthase (NOS) (58), and this effect is inhibited by cytokines important in the pathogenesis of IR (59). Importantly, the effect of insulin on NOS is mediated through the same intracellular signaling pathway as the effect of insulin on glucose metabolism (60). Thus, IR in glucose metabolism and in the vasculature can be explained on the basis of a single defect. Recent data suggest that the metabolic and vascular actions of insulin are closely linked. Insulinresistant states exhibit diminished insulin-mediated glucose uptake into peripheral tissues as well as impaired insulinmediated vasodilation and impaired endothelium-dependent vasodilation to the muscarinic receptor agonist, acetylcholine. Thus, insulin action in peripheral tissues is probably linked to its action on endothelium. Basal forearm blood flow in diabetic and nondiabetic subjects is comparable. The forearm blood flow responses to both methacholine chloride and nitroprusside are significantly attenuated in diabetic compared with nondiabetic subjects. Endothelial dysfunction is also detectable in young normotensive first-degree relatives of subjects with type 2 diabetes (61). There is a significant association between endothelial dysfunction and IR in young relatives of diabetic subjects independent of the classic cardiovascular risk factors (62). Caballero et al. (61) studied vascular reactivity in both the micro- and macrocirculation as well as biochemical markers of endothelial function in four age- and sex-comparable groups: healthy normoglycemic subjects with no family history of type 2 diabetes (controls), healthy normoglycemic subjects with a history of type 2 diabetes in one or both parents (relatives), subjects with impaired glucose tolerance (IGT), and patients with type 2 diabetes without vascular complications (diabetes). The vasodilatory responses to acetylcholine chloride were reduced in relatives, IGT, and diabetes compared with controls, as were the responses in the brachial artery diameter during reactive hyperemia. Compared with control subjects, endothelin-1 was significantly higher in all groups, vwf was higher only in the diabetic group, and soluble ICAM levels were higher in the IGT and diabetic groups, whereas soluble VCAM concentrations were higher in the relatives and those with diabetes. These results suggest that abnormalities in vascular reactivity and biochemical markers of endothelial cell activation are present early in individuals at risk of developing type 2 diabetes, even at a stage when normal glucose tolerance exists, and that factors in addition to IR, such as genetic factors, may be operative. B. Proposed mechanisms linking endothelial dysfunction with CVD in diabetes As discussed above, insulin itself has vasodilatory actions via a NO-dependent mechanism (63). In healthy subjects, insulin dilates arterioles supplying skeletal muscle, probably through enhancement of NO production. Some in vitro studies have documented that insulin regulates NOS, the enzyme that synthesizes NO from arginine. This action may be impaired in insulin-resistant subjects, an abnormality that might be attributable to either impairment in the ability of the endothelium to produce NO or enhanced inactivation of NO (63). Because NO plays a critical role in the maintenance of vascular health (41), this abnormality may explain much of the increased CVD in the IRS. Impairment of insulin action on glucose metabolism assessed by glucose clamp parallels impairment of insulin action on the vasculature. Thus, obesity and type 2 diabetes are associated with resistance to the vascular effects of insulin. As discussed above, abnormalities in insulin signaling lead not only to IR in glucose metabolism but also abnormalities in the vasculature (39). Recent data suggest that insulin signaling through the phosphatidylinositol 3-kinase pathway is important in NO production in human vascular endothelial cells (64 66). Disruption of this signaling pathway (which is known to be associated with IR), may lead to a disturbance in NO production and contribute to vascular disease.

7 Fonseca et al. Cardiovascular Risk Factors in Diabetes Endocrine Reviews, February 2004, 25(1): Increased levels of asymmetric dimethylarginine (ADMA) are associated with endothelial dysfunction and increased risk of CVD. ADMA is an endogenous and competitive inhibitor of NOS (67). Plasma levels of this inhibitor are elevated in patients with atherosclerosis and in those with risk factors for atherosclerosis (67). In these patients, plasma ADMA levels are correlated with the severity of endothelial dysfunction and atherosclerosis. By inhibiting the production of NO, ADMA may impair blood flow, accelerate atherogenesis, and interfere with angiogenesis. Thus, plasma ADMA may be a novel risk factor for vascular disease (67). Stuhlinger et al. (68) have demonstrated that plasma ADMA concentrations are positively correlated with impairment of insulin-mediated glucose disposal in subjects with the metabolic syndrome, independent of hypertension. Pharmacological intervention with rosiglitazone enhanced SI and reduced ADMA levels (68). Increases in plasma ADMA concentrations may thus contribute to the endothelial dysfunction observed in insulin-resistant patients and also contribute to CVD in diabetes. Tetrahydrobiopterin (BH4), an essential cofactor for the catalytic activity of enos, is depleted during states of oxidative stress because of excessive oxidation (69). Depleted BH4 causes enos to uncouple, which results in decreased NO production. IR has been shown to diminish the activity of the enzyme that produces BH4 in human coronary arteries with resultant BH4 depletion and endothelial dysfunction, which is reversed by BH4 administration (70). Treatment with BH4 has been shown to improve endothelial function in experimental diabetes (71). Disturbances in other functions of the endothelium, such as increased expression of adhesion molecules and suppression of inflammation (discussed in Section IV), also play an important role in diabetes. In summary, endothelial dysfunction is the earliest abnormality associated with CVD and occurs very frequently in patients with diabetes, often preceding the onset of hyperglycemia. IV. Impaired Fibrinolysis and Prothrombotic State A. Epidemiology and clinical observations Factors contributing to a prothrombotic state in diabetes are summarized in Table 3. The endogenous fibrinolytic system represents equilibrium between activators of plasminogen (primarily tissue type plasminogen activator) and inhibitors of these activators [such as plasminogen activator inhibitor type 1 (PAI-1) (72)]. Low-grade coagulation is a continuous process, and thus the fibrinolytic activity is necessary to maintain the fluidity of blood. Excessive inhibition of fibrinolysis will lead to coagulation and thrombosis, a critical process in cardiovascular events. Impaired fibrinolytic function in diabetes correlates with severity of vascular disease in diabetes and is a risk factor for myocardial infarction in both diabetic and nondiabetic subjects (73 77). Impaired fibrinolysis is now recognized as being an important component of the IRS and probably contributes considerably to the increased risk of cardiovascular events (30, TABLE 3. Impact of IR and diabetes on thrombosis and fibrinolysis Factors predisposing to thrombosis 1 Platelet hyperaggregability 1 Platelet camp and cgmp 1 Thromboxane synthesis Elevated concentrations of procoagulants 1 Fibrinogen 1 vwf and procoagulant activity 1 Thrombin activity Decreased concentration and activity of antithrombotic factors 2 Antithrombin III activity Factors attenuating fibrinolysis Decreased tpa activity Increased PAI-1 synthesis and activity Increased blood viscosity tpa, Tissue type plasminogen activator; 1, increased; 2, decreased. 72). Plasma PAI-1 antigen and activity are elevated in a wide variety of insulin-resistant subjects including obese subjects with and without diabetes and women with the polycystic ovarian syndrome. Elevated levels of fasting insulin are associated with impaired fibrinolysis and hypercoagulability in subjects with normal glucose tolerance (78). Hyperinsulinemia is associated primarily with impaired fibrinolysis in subjects with glucose intolerance. Excess risk for CVD associated with hyperinsulinemia and glucose intolerance may be mediated in part by enhanced potential for acute thrombosis. Finally, abdominal fat produces PAI-1 and could contribute to increased plasma PAI-1 concentrations in human obesity associated with IR (79). Coagulation disorders also play a role in increasing the risk of CHD in patients with type 2 diabetes (80). Platelets from patients with diabetes are more sensitive to several aggregating agents and have increased numbers of glycoprotein receptors and a lower activity of guanylate cyclase (80, 81). These factors may contribute to the documented hyperreactivity of platelets in patients with type 2 diabetes. Other factors in patients with type 2 diabetes include alterations in serum fibrinogen, and factors V, II, and VII, which have all been linked to the risk of myocardial infarction. Increased D-dimer, vwf antigen, A-II antiplasmin, and decreased antithrombin III have also been reported in patients with type 2 diabetes and reviewed elsewhere (80). Many of these abnormalities are nonspecific, and the association of IR with coagulation abnormalities is less robust than that with abnormal fibrinolysis. Serum fibrinogen is an acute phase reactant, and elevated plasma fibrinogen may be a manifestation of inflammation like CRP. Nevertheless, coagulation abnormalities probably play a role in increasing the frequency and severity of thrombotic events in patients with diabetes. B. Proposed mechanisms linking impaired fibrinolysis and thrombosis with CVD in diabetes Probably the most important component of disturbed coagulation in diabetes relates to the abnormal fibrinolysis due to changes in the dynamic equilibrium between endogenous tissue plasminogen activator and PAI-1 (82). Insulin, proinsulin, VLDL cholesterol, and various cytokines regulate PAI-1 synthesis and release. The greatest elevations in PAI-1

8 160 Endocrine Reviews, February 2004, 25(1): Fonseca et al. Cardiovascular Risk Factors in Diabetes occur when there is a combination of hyperinsulinemia, hyperglycemia, and increased FFA in obese insulin-resistant subjects (83), and impaired fibrinolysis is closely related to the metabolic syndrome (84). Basal fibrinolytic activity is decreased in patients with type 2 diabetes; this may accelerate atherosclerosis by exposing vascular luminal wall surfaces to persistent and recurrent thrombi. There is also evidence that PAI-1 content is increased in atherosclerotic lesions of patients with type 2 diabetes (85), suggesting that interventions to reduce IR and improve glycemic control may improve the fibrinolytic response. Diabetes is associated with increased PAI-1 in the arterial wall, which could decrease local fibrinolysis and elevate thrombus formation and the unfavorable evolution of atherosclerotic plaques (86). Insulin also inhibits platelet function, PAI-1, and transcription factors associated with coagulation (87 89). Dysregulation of these actions may contribute to the hypercoagulability in insulin-resistant states. Immunohistochemical analysis of coronary lesions from patients with coronary artery disease has demonstrated an imbalance of the local fibrinolytic system with increased coronary artery tissue PAI-1 in patients with type 2 diabetes (85). In summary, increased PAI-1 and, to a lesser extent, increased coagulation are very closely linked to IR and thereby could contribute to CVD in diabetes. V. Inflammation and CVD in Diabetes A. Epidemiology and clinical observations Inflammation has recently been associated with cardiovascular events in several studies, and the relative ease of measurement of low-grade inflammation by high sensitive FIG. 2. Schematic representation of the interactions between inflammation, IR, and atherosclerosis. (hs)-crp, makes it attractive to use this risk marker in clinical practice (90 93). None of the studies describing an association between inflammation and cardiovascular events have focused exclusively on diabetics, although diabetics were included in most of these studies. However, the recent finding that inflammatory markers also predict the development of incident diabetes has led to the formulation of the hypothesis that inflammation may be the underlying link between diabetes and CVD. Several markers of inflammation have been used in various studies. Data suggest that plasma CRP, measured using an hs-crp assay, appears to have a significant predictive value in determining the risk of future coronary events (90). CRP is an acute-phase protein produced by the liver in response to cytokine production (IL-6, IL-1, TNF ). Figure 2 summarizes the hypothetical link between cytokines secreted by adipose tissue (adipokines), inflammation, and CVD in diabetes. The standard CRP test determines levels of CRP that increase up to 1000-fold in response to infection or tissue destruction but cannot adequately assess the normal range and thus cannot be used for cardiovascular risk prediction. hs-crp assays detect changes in levels of CRP within the normal range. High levels in this range have been proven to predict future cardiovascular events. Other inflammatory risk factors, including oxidized lipids, infectious agents, and cytokine produced from adipocytes or other inflammatory cells, stimulate production of IL-6, which serves as a messenger cytokine that stimulates the liver to produce inflammatory substances such as CRP (93). Several epidemiological studies have documented that an elevated hs-crp with a good predictor of the development of coronary artery disease is a strong predictor of future coronary events in apparently healthy subjects and has prog-

9 Fonseca et al. Cardiovascular Risk Factors in Diabetes Endocrine Reviews, February 2004, 25(1): nostic value in patients for acute coronary syndromes. In the Physician s Health Study, the baseline plasma concentration of CRP predicted the risk of future myocardial infarction and stroke (94). Moreover, the reduction associated with the use of aspirin in the risk of a first myocardial infarction appears to be directly related to the level of CRP, raising the possibility that antiinflammatory agents may have clinical benefits in preventing CVD. CRP was measured in baseline blood samples from 122 apparently healthy participants in the Women s Health Study who subsequently suffered a first cardiovascular event and from 244 age- and smokingmatched control subjects who remained free of CVD during a 3-yr follow-up period (90, 95). Women who developed cardiovascular events had higher baseline CRP levels than control subjects, such that those with the highest levels at baseline had a 5-fold increase in risk of any vascular event and a 7-fold increase in risk of myocardial infarction or stroke. Risk estimates were independent of other risk factors, and prediction models that included CRP provided a better method to predict risk than models that excluded CRP. Several other prospective studies have been published in which hs-crp has predicted cardiovascular events after correction for various other risk factors (95 101). In all of these studies, the subjects have been free of clinical CVD at baseline. In 1997, Ridker et al. (94) demonstrated that subjects in the highest quartile of CRP ( 2.1 mg/liter) had more than 2.5 increased risk of CVD compared with those in the lowest quartile of CRP ( 0.55 mg/liter). Importantly, an elevated hs-crp is specific for cardiovascular events but not for other diseases such as cancer (95, 102). Recently, Ridker et al. (95) demonstrated that hs-crp measurements add to the predicted value of total cholesterol in determining the risk of a first myocardial infraction. More importantly, subjects who have hs-crp greater than the 75th percentile and a total cholesterol less than 75th percentile had a significantly increased risk of events. The impact of hs-crp interaction with lipids on cardiovascular risk has been reviewed recently (103). It appears that the predicted value of hs-crp is independent of the total cholesterol and is additive to the risk determined by total cholesterol. It is particularly important to compare the value of measuring plasma CRP protein and LDL cholesterol levels in predicting cardiovascular risk. Ridker et al. (104) measured both at baseline in 27,939 apparently healthy American women, who were then followed for a mean of 8 yr. Although CRP and LDL cholesterol were minimally correlated (r 0.08), baseline levels of each had a strong linear relation with the incidence of cardiovascular events. After adjustment for other factors, the relative risks of first cardiovascular events according to increasing quintiles of CRP, as compared with the women in the lowest quintile, were 1.4, 1.6, 2.0, and 2.3 (P 0.001), whereas the corresponding relative risks in increasing quintiles of LDL cholesterol, as compared with the lowest, were 0.9, 1.1, 1.3, and 1.5 (P 0.001) (104). A significant proportion of events occurred among women with normal LDL cholesterol. By contrast, because CRP and LDL cholesterol measurements tended to identify different highrisk groups, screening for both biological markers provided better prognostic information than screening for either alone. Independent effects were also observed for CRP in analyses adjusted for all components of the Framingham risk score (104). These data suggest that the CRP level is a stronger predictor of cardiovascular events than the LDL cholesterol level and that it adds prognostic information to that conveyed by the Framingham risk score. Because many of the features of the metabolic syndrome are associated with increased levels of CRP, it is important to evaluate the relationship between the two (105). Among 14,719 apparently healthy women who were followed up for an 8-yr period, 24% had the metabolic syndrome at study entry. At baseline, median CRP levels for those with zero, one, two, three, four, or five characteristics of the metabolic syndrome were 0.68, 1.09, 1.93, 3.01, 3.88, and 5.75 mg/liter, respectively [P(trend) ] (106). Over the 8-yr followup, cardiovascular event-free survival rates based on CRP levels above or below 3.0 mg/liter were similar to survival rates based on having three or more characteristics of the metabolic syndrome. At all levels of severity of the metabolic syndrome, however, CRP added prognostic information on subsequent risk (106). Recently, recommendations have been made on standardization of the hs-crp assay methodology and simplification of ranges of plasma hs-crp as risk indicators. Using widely available high-sensitivity assays, CRP levels of less than 1, 1 3, and more than 3 mg/liter correspond to low-, moderate-, and high-risk groups for future cardiovascular events (107). Individuals with LDL cholesterol below 130 mg/dl who have CRP levels more than 3 mg/liter represent a high-risk group often missed in clinical practice. The addition of CRP to standard cholesterol evaluation may thus provide a simple and inexpensive method by which to improve global risk prediction and compliance with preventive approaches (107). In summary, hs-crp measurement has shown to be an important risk marker of CVD, as well as the risk of developing diabetes, and thereby has several potential clinical applications. Plasma hs-crp measurements could serve as an adjunct to lipid screening in the protection of individuals at high risk for both conditions, e.g., the metabolic syndrome. It may provide us with a method to better target statin therapy, particularly in the setting of primary prevention. It may have potential prognostic value in acute coronary syndrome. Finally, inflammation is likely to represent a new target for treatment and prevention of cardiovascular events and possibly even diabetes. B. Proposed mechanisms linking inflammation with CVD in diabetes The mechanism linking inflammation with CVD in diabetes is not clear. It has been proposed that the markers of inflammation may be somewhat nonspecific. Nevertheless, a recent paper by Burke et al. (108) reported that CRP is present in the lipid core of atherothrombotic lesions in subjects who die suddenly. The CRP is present adjacent to cholesterol and clots. In this context, a recent in vivo study has demonstrated that CRP, at concentrations known to predict adverse vascular events, directly quenches the production of the NO, in part, through posttranscriptional effects on enos mrna stability (95 97, 101, 109). CRP promotes monocyte chemo-