Primary Prevention of Cardiovascular Disease in People With Dysglycemia

The first indication that hyperglycemia may represent a cardiovascular risk factor dates back to 1965 when the Bedford Study (20) and the Tecumseh Study (21) were published. These cross-sectional studies indicated that increased blood glucose levels were associated with greater cardiovascular mortality. However, it was only 15 years later that the first systematic analysis of this relationship appeared. The International Collaborative Group reviewed findings from 11 different studies with an observation period of 4–15 years (22). However, this analysis did not yield univocal results, and it was not possible to support a clear-cut association between asymptomatic hyperglycemia and CVD (22). A close association between fasting plasma glucose or A1C levels and CVD was found during a 10-year follow-up in the Uusitupa study, which included 133 newly diagnosed diabetic subjects (22).

Coutinho et al. (17) carried out a meta-analysis of 20 studies for a total of 95,783 nondiabetic subjects who had, during a mean follow-up of 12.4 years, 3,707 cardiovascular events. In this meta-analysis, an exponential relationship between the risk of cardiovascular events and both fasting and postload plasma glucose levels was found. Such a relationship extended below diagnostic blood glucose levels for IFG or IGT. The Cardiovascular Health Study (23) assessed cardiovascular risk factors in people aged ≥65 years and showed that CVD was related to both fasting and postchallenge glucose. Even if much research has been performed in Western populations, recent evidence demonstrated similarly increased risks in Asian populations. Analysis of cohorts from the Asia Pacific region indicates a positive continuous association between usual blood glucose and cardiovascular risk. This association holds for glucose levels as low as 4.9 mmol/l, i.e., below diagnostic fasting plasma glucose for diabetes and IGT (24). Evidence has been recently published to support that higher-than-optimum blood glucose is a major cause of cardiovascular mortality in most world regions (25). While diabetes was associated with a total of 959,000 deaths, 1,490,000 deaths for ischemic heart disease and 709,000 from stroke were attributable to higher-than-optimum blood glucose, accounting for 21 and 13% of all deaths from those conditions.

An even greater cardiovascular risk is likely to be associated with excessive postload glucose levels (18,26–29). The DECODE study (26) evaluated data from 13 epidemiologic studies and found that the mortality risk of IFG subjects was 1.2, while it was 1.5 in IGT individuals. Subsequent analysis of the DECODE database explored the association between categories of glucose tolerance and ischemic heart disease, CVD, and mortality for all causes (18). Upon adjustment for BMI, blood pressure, total cholesterol, and smoking habit, relative mortality risk did not increase as a function of fasting plasma glucose levels, while it remained highly significant for postload glucose values. Moreover, when excess mortality was considered, no association was found with isolated fasting hyperglycemia, whereas IGT and IFG+IGT accounted for 44 and 14% of the excess mortality, respectively. Similarly, in the Funagata Diabetes Study (27), cardiovascular mortality in IGT was very close to that of subjects with diabetes and much greater that that in IFG individuals. Among 3,092 adults aged 30–74 years (28), an increased relative risk of all-cause mortality of 1.6 (95% CI 1.0–2.6) was found in subjects with newly diagnosed diabetes based on isolated postchallenge hyperglycemia (glucose ≥11.1 mmol/l), whereas the relative risk of death in those with known diabetes was 2.1 (95% CI 1.4–3.2). Mortality was also increased in IGT subjects who had a relative risk of 1.3 (95% CI 1.0–1.6). In the Hoorn Study (29), the relative risk of cardiovascular mortality started to be significant from a fasting glucose level of 6.1 mmol/l onward. Both postloading glucose and A1C showed a linear correlation to cardiovascular risk, even in the nondiabetic range. However, the authors of this study recently showed that the association between IFG and the risk of CVD mortality depends on the conversion to diabetes during the follow-up (30). In fact, subjects who converted from IFG to diabetes had a more than twofold risk of CVD mortality (IFG 6.1: 2.47 [1.17–5.19]; IFG 5.6: 2.14 [1.12–4.10]) than subjects with normal fasting glucose, whereas IFG subjects who did not develop diabetes did not have significantly higher CVD mortality risks (IFG 6.1: 1.50 [0.72–3.15]; IFG 5.6: 1.15 [0.69–1.93]).

A1C level provides an integrated measure of glycemic control, including both fasting and postprandial glucose levels. A meta-analysis of observational studies has documented increased risk of CVD associated with higher levels of A1C in diabetic patients (31).

Unfortunately, few longitudinal studies are available in nondiabetic subjects. In a study carried out in the U.K. including ∼5,000 individuals with A1C >5.0%, a direct correlation was found with total and coronary heart disease mortality (15), with a 41% increase in the relative risk per percent point increase in A1C. Similar results have been produced in a Dutch study including 2,363 nondiabetic individuals, in whom the mortality risk increased by 51% per 1.4% point of A1C (29). In the Rancho Bernardo Study (32), an association between A1C and cardiovascular mortality was found among women but not among men. The association between A1C levels and peripheral arterial disease has been recently analyzed in the National Health and Nutrition Examination Survey (NHANES) 1999–2002 (33). The study confirmed that A1C was a significant predictor of peripheral arterial disease, an association that persisted even within the normal range and after multivariable adjustment.

Whether hyperglycemia is the culprit of such an elevated cardiovascular risk remains to be elucidated, since it may be a simple but very useful marker for a condition characterized by concomitance of multiple cardiovascular risk factors. Participants in the San Antonio Heart Study who did not have diabetes at baseline but developed type 2 diabetes during an 8-year follow-up had higher total and LDL cholesterol, triglycerides, BMI, and blood pressure measurements and lower HDL cholesterol than subjects who did not develop type 2 diabetes (34). Based on these observations, the “ticking clock” hypothesis was formulated, which postulates that “the clock for coronary heart disease starts ticking before the onset of clinical diabetes” (34). The findings of the Nurses’ Health Study (35) provide support to this hypothesis, showing that the risk for CVD occur long before the onset of overt hyperglycemia. In this large prospective cohort study, an elevated risk of myocardial infarction and stroke was found before clinical diagnosis of type 2 diabetes, compared with women who remained nondiabetic throughout the study. Increased risk persisted in obese and nonobese women as well as in those with or without a family history of diabetes (35). Cardiovascular risk was already increased 15 years before diagnosis of diabetes and increased further upon clinical diagnosis of the disease. In that study, women converting to type 2 diabetes had a more atherogenic risk profile (e.g., higher BMI values and greater prevalence of hypertension and hypercholesterolemia) than individuals retaining normal glucose tolerance during the study (35).

In the Botnia Study, ∼50% of the subjects classified as IFG and/or IGT also met the diagnosis of metabolic syndrome, and these individuals had greater risk for both coronary heart disease and stroke (36). This finding is in agreement with the results of a pooled analysis of several prospective studies published from 1988 to 2004 indicating that the relative risk for CVD in individuals with metabolic syndrome, as defined by National Cholesterol Education Program Adult Treatment Panel III, was 1.65 (95% CI 1.38–1.99) and that the relative risks for all-cause mortality and diabetes were 1.27 (0.90–1.78) and 2.99 (1.96–4.57), respectively (37).

Pre-diabetic dysglycemia acts in concert with other cardiovascular risk factors within the picture of the metabolic syndrome. Subjects with dysglycemia often have insulin resistance and hyperinsulinemia, elevated blood pressure, high total and VLDL triglycerides, low HDL cholesterol, small and dense LDL, central adiposity, elevated levels of plasminogen activator inhibitor 1, and low-grade inflammation (38). In spite of all this, experimental evidence supports a direct effect of mild alteration of glucose homeostasis on the atherogenic process. Early alterations of glucose homeostasis is often heralded by postprandial hyperglycemia. Higher and persistent elevation of postprandial plasma glucose levels contribute significantly to overall glycemic exposure of body tissues (39). Moreover, excessive and repeated glucose excursions lead to increased glycemic variability. Overall, glycemic exposure and glycemic variability can both contribute to favor development of atherosclerosis. The arterial wall is indeed a critical target for such glycemic insult. Protein glycosylation causes cross-linking of collagen and other extracellular matrix proteins in the arterial wall, increasing susceptibility to atherosclerosis. Glycosylation of LDL results in more atherogenic particles, macrophage activation, and foam cell formation. A more direct action is exerted on endothelial cells leading to endothelial dysfunction, an early manifestation of atherosclerosis, and a strong predictor for CVD (40).

In type 2 diabetes, glucose fluctuations during postprandial periods and, more generally, during swings exhibited a more specific triggering effect on oxidative stress than chronic hyperglycemia (41). Several in vitro studies have demonstrated increased expression of markers of oxidative stress in cells exposed to fluctuating glucose concentrations (42). Hyperglycemia in non–insulin-dependent tissues such as endothelium results in overproduction of superoxide and free radicals by the mitochondrial electron transport chain. Free radicals may mediate some of the effects associate with hyperglycemia: vasoconstriction, activation of coagulation, and increased adhesion molecule expression. As described by Brownlee (43), NADPH inhibition by superoxide slows down the glycolytic flux and leads to accumulation of glycolytic precursors, which in turn allow activation of pathways that normally account for a small proportion of intracellular glucose metabolism: the polyoil pathway, the hexosamine pathway, protein kinase C activation, and the advanced glycation end product formation. Chronic and acute hyperglycemia alter the activity of protein kinase C (44), a key enzyme in signal transduction. Its activation increases secretion of endothelin, type IV collagen, and fibronectin; enhances expression of adhesion molecules in endothelial cells; and activates macrophage migration processes.

Postprandial hyperglycemia and glycemic instability are also associated with a proinflammatory state, as indicated by generation of advanced glycation end products, increased C-reactive protein, and tumor necrosis factor-α. Acute hyperglycemia also activates carbonyl stress, leading to formation of highly reactive species such as 3-deoxyglucosone and methylglyoxal (45).

In summary, epidemiological data linking dysglycemia with cardiovascular risk are supported by experimental findings illustrating the mechanisms through which even mild alterations in glucose homeostasis may favor development and progression of the atherosclerotic lesion.

Final evidence for glucose as a cardiovascular risk factor can be based only on intervention studies. The heated discussion about the 15% reduction of myocardial infarction in the UKPDS has not yet settled (11). Still, epidemiologic analysis of those data has indicated a 14% risk reduction for myocardial infarction for each 1% reduction in A1C (11). These results have received further support from recent work in type 1 diabetic patients. In the follow-up report of the Diabetes Control and Complications Trial, reduced carotid intima-media thickness was found in intensively treated patients compared with those on conventional insulin therapy (46). Reduced progression of atherosclerosis was then confirmed by a more recent publication reporting a 42% reduction of the risk of any predefined CVD outcome (P = 0.02) and 57% reduction of the risk of the first occurrence of nonfatal myocardial infarction, stroke, or death from CVD (P = 0.02) (47).

If strict glycemic control is recommended to reduce risk for micro- and macrovascular complications, a large number of individuals still present signs of cardiovascular involvement at the time of diabetes diagnosis. Therefore, effort should be made to reduce cardiovascular risk at the time when dysglycemia develops rather than when hyperglycemia develops. A few years ago, the Third Report of the National Cholesterol Education Program Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) (48) identified diabetes as a coronary heart disease risk equivalent. If that is the case, prevention of diabetes should be pursued with the same effort paid to reduce other known cardiovascular risk factors. Direct evidence for such an opportunity is still scanty, but all diabetes prevention trials indicate a positive effect on cardiovascular risk profile.

In the Finnish Diabetes Prevention Study (49), lifestyle modification in IGT patients reduced conversion to diabetes by 58%. Along with lower risk of diabetes, almost all cardiovascular risk factors were significantly reduced. At the end of the 3.2-year follow up, diastolic and systolic blood pressure, serum triglycerides, and plasma insulin levels were all much lower in the intervention group than in control individuals (49). The results of the Diabetes Prevention Program (50) provide further evidence of the potential benefit of lifestyle changes beyond diabetes prevention. Hypertension was present in 30% of participants at study entry and then increased in the placebo and metformin groups, although it significantly decreased with intensive lifestyle intervention. Triglyceride levels fell in all treatment groups, but fell significantly more with intensive lifestyle intervention. Total cholesterol and LDL cholesterol levels were similar among treatment groups. Intensive lifestyle intervention significantly increased the HDL cholesterol level and reduced the cumulative incidence of the proatherogenic LDL phenotype B. After a 3-year follow-up, the use of pharmacologic therapy to achieve established goals in the intensive lifestyle group was 27–28% less for hypertension and 25% less for hyperlipidemia compared with placebo and metformin groups. Lifestyle intervention also reduced levels of nontraditional cardiovascular risk factors, such as C-reactive protein, relative to both placebo and to a lesser degree to metformin (51). Finally, incidence of the metabolic syndrome was reduced by 41% in the lifestyle group (P < 0.001) and by 17% in the metformin group (P < 0.03) compared with placebo (52).

Though lifestyle modification appears to be highly effective, use of medications for prevention of or early treatment of dysglycemia also results in the improvement of cardiovascular risk factors, as indicated by the use of metformin in the Diabetes Prevention Program. Anti-obesity drugs such as orlistat, an inhibitor of pancreatic and gastrointestinal lipases, have been shown to reduce conversion toward diabetes in obese individuals irrespective of their glucose tolerance. In the XENDOS (Xenical in the Prevention of Diabetes in Obese Subjects) trial (n = 3,305 obese nondiabetic individuals), orlistat, combined with lifestyle changes, reduced the incidence of type 2 diabetes compared with placebo (plus lifestyle changes) by 37.3% (P = 0.0032) (54). In those individuals, a 7-kg body weight reduction was associated with lower levels of LDL cholesterol and greater reduction blood pressure values as compared with control subjects (54). Rimonabant, a cannabinoid receptor (CB1) antagonist, has generated even greater expectation. The RIO (Rimonabant in Obesity) studies have shown that rimonabant, at the dose of 20 mg/day, causes >5% weight loss in 30–40% of patients and >10% in 10–20% of them compared with diet alone over a 2-year period, as well as an extra 8–10% increase in HDL cholesterol and a 10–30% reduction in triglycerides. In patients with type 2 diabetes, the treatment results in improvement in insulin resistance, glycemic control, adipokines, and inflammatory proteins (54–56).

A 62% reduction in diabetes and greater conversion to normal glucose tolerance has been documented in the DREAM (Diabetes Reduction Assessment With Ramipril and Rosiglitazone Medication) study in IGT subjects treated with 8 mg rosiglitazone compared with the control group. Active treatment was also associated with modest though highly significant reduction in diastolic and systolic blood pressure (57). No effect was found in terms of cardiovascular events with the exception of increased risk for heart failure in people on rosiglitazone. However, none of the prevention trials so far discussed was powered nor designed to detect an effect on cardiovascular events. Still, supportive observation is available. The Malmö Prevention Study was carried out 30 years ago (58). In that study, 147 pre-diabetic individuals were randomly treated with diet and placebo, diet and tolbutamide, or no therapy at all. As expected, diet was effective in reducing conversion to diabetes when compared with no treatment. However, even more effective was the concomitant use of tolbutamide, with not a single case of diabetes observed over a 10-year follow-up. Fifteen years after the publication of those results, data of a longer follow-up were reported showing significant reduction in all-cause mortality and ischemic heart disease in individuals initially treated with tolbutamide (60). The study is questionable because of small sample size. The NAVIGATOR study will test a similar hypothesis in >9,000 dysglycemic individuals by targeting postprandial glucose with nateglinide (61).

The results of the Study to Prevent Non-Insulin-Dependent Diabetes Mellitus (STOP-NIDDM) (62) seem to confirm the role for specific treatment of postprandial glucose levels, at least in individuals who still have a moderate increase in fasting plasma glucose concentration. In IGT patients, treatment with acarbose, an intestinal α-glucosidase inhibitor, not only reduced conversion to type 2 diabetes by 30%, but was also associated with reduction of the risk of any cardiovascular event by 49%, of an acute myocardial infarction by 91%, and of developing hypertension by 34%. Still lacking is the evidence that strict maintenance of normal fasting plasma glucose levels also may reduce cardiovascular risk. The results of the ORIGIN study may provide an initial answer to the question (63).

In summary, epidemiologic data support the hypothesis of a direct and independent relationship between blood glucose levels and CVD. The lack of a clear-cut threshold value in diabetic patients and the persistence of the relationship in the nondiabetic population suggest that blood glucose is a continuous variable, similarly to other CVD risk factors. Prevention or delay in onset of diabetes is of utmost importance because chronic hyperglycemia is associated with CVD, and the risk starts to increase long before the onset of clinical diabetes.

Moreover, dysglycemia is often associated with other risk cardiovascular factors to result in the metabolic syndrome, a condition loaded with high propensity to atherosclerosis. Reducing the burden of CVD will then require identification of initial alteration of glucose homeostasis, as well as careful search of any associated CV risk factor. Lifestyle modification has been shown to be quite effective in preventing conversion of IGT patients toward overt diabetes, also resulting in effective improvement of many cardiovascular risk factors. Pharmacological intervention may provide benefit as well, although further studies are needed to characterize the benefits of these therapeutic strategies to reduce cardiovascular risk. In the meantime, clinicians should evaluate and treat all cardiovascular risk factors (64), including dysglycemia.