Diabetes mellitus is a major independent risk factor for cardiovascular disease.
The worldwide prevalence of diabetes is increasing, driven primarily by the rise in type 2 diabetes.
The pathophysiology of diabetic cardiovascular disease is multifactorial and incompletely understood.
Lifestyle intervention is fundamental for the prevention of both type 2 diabetes and its cardiovascular complications.
Glucose control is important for the management of diabetes, but the most appropriate strategies and glucose targets for cardiovascular disease prevention are still uncertain.
In addition to lifestyle changes, pharmacologic interventions to treat hypertension and dyslipidemia in diabetes are essential for cardiovascular disease prevention.
The use of antiplatelet interventions to prevent cardiovascular disease is still controversial but recommended in most patients with diabetes.
The incidence and prevalence of diabetes mellitus are on the rise in the United States and globally, almost entirely due to the growing pandemic of type 2 diabetes. Given that diabetes is a major independent risk factor for cardiovascular disease (CVD), in some clinical contexts considered a coronary disease equivalent, the prevention of diabetes and the management of its associated CVD risk factors are of paramount public health importance. This chapter reviews the epidemiology and preventive strategies for diabetes and its associated CVD complications, with special emphasis on type 2 diabetes, which accounts for more than 90% of diabetes cases worldwide.
Epidemiology of Diabetes Mellitus
Diabetes mellitus is a group of diseases characterized by insufficient production of insulin or by the failure of the body to appropriately respond to insulin, resulting in hyperglycemia. Vascular complications, the principal clinical risk associated with diabetes, are classified as microvascular (diabetic retinopathy, nephropathy, neuropathy) and macrovascular (ischemic heart disease, cerebrovascular disease, peripheral vascular disease).
The World Health Organization and the American Diabetes Association (ADA) criteria for the diagnosis of diabetes have evolved during recent decades, summarized in Table 21-1 . A diagnosis of diabetes is usually based on tests repeated on at least two different days, unless hyperglycemia is unequivocal or the person is symptomatic.
|2010 ADA |
|1999 WHO |
|1997 ADA||FPG ≥7.0 mmol/L (126 mg/dL)|
|1985 WHO |
1979 NDDG *
Approximately 90% or more of cases of diabetes mellitus are characterized by relative insulin deficiency with a backdrop of insulin resistance and are classified as type 2 diabetes mellitus. The etiology of type 2 diabetes is multifactorial, encompassing genetic, environmental, and behavioral factors, but the exact mechanistic underpinning has not yet been determined; a number of predisposing factors for the development of type 2 diabetes are summarized in Table 21-2 .
|Prediabetes, defined as impaired glucose tolerance (2-hr plasma glucose concentration, 140 mg/dL [7.8 mmol/L] to 199 mg/dL [11.0 mmol/L] during an oral glucose tolerance test) or impaired fasting glucose (fasting plasma glucose concentration, 100 mg/dL [5.6 mmol/L] to 125 mg/dL [6.9 mmol/L])|
|Obesity (BMI ≥30 kg/m 2 )|
Type 1 diabetes mellitus results from primary beta-cell loss leading to absolute insulin deficiency, representing less than 10% of cases of diabetes mellitus. The etiology of type 1 diabetes is also multifactorial and poorly understood, although an autoimmune component has been implicated in most cases. Other forms of diabetes not covered in this chapter are gestational diabetes and numerous less common causes (e.g., monogenic defects in insulin production or action; diabetes secondary to other pathologic conditions of the pancreas, such as pancreatitis or tumors; and drugs or chemicals causing beta-cell toxicity).
Incidence and Prevalence of Diabetes Mellitus
More than 180 million people worldwide were estimated by the World Health Organization to have diabetes mellitus in 2008, increasing from an estimated 135 million in 1995 and projected to rise to 366 million by 2030 ( Fig. 21-1 ). This represents a rise in prevalence adjusted for population growth from 2.8% in 2000 projected to 4.4% in 2030. These numbers probably underestimate the burden of diabetes mellitus in the developing world, where only 25% to 30% of cases are diagnosed.
The 2007 diabetes mellitus prevalence estimates for the United States included 17.9 million people with diagnosed diabetes and 5.7 million more undiagnosed, representing 7.8% of the U.S. population. The number of Americans with diagnosed diabetes is projected to increase to 29 million by 2050, with less than one third of this increase attributable to population growth. More than 1.6 million new cases of diabetes mellitus were diagnosed in the United States in 2007 alone, representing a marked increase from 625,000 new cases annually in 1990-1992, even with adjustment for total population (incidence rates 5.3 versus 2.4 per 1000 people per year).
A number of populations are especially vulnerable to development of type 2 diabetes. Older persons, women, and especially elderly women are particularly susceptible ( Fig. 21-2 ). In 2002 in the United States, 1.7% of the population aged 20 to 39 years had diagnosed diabetes compared with 15.1% of those ≥60 years. Diabetes is also more prevalent in certain ethnic groups, including African Americans, Asian Indians, Hispanic or Latino Americans, Native Americans, and Pacific Islanders, among others. For example, the age- and sex-standardized prevalence of diagnosed diabetes reported in NHANES 1999-2002 was almost double in African Americans and Hispanic or Latino Americans compared with non-Hispanic white Americans. The prevalence of diabetes is high but variable in different Native American populations, reaching 40% to 50% in Pima Indians older than 35 years. Somewhat paradoxically, higher socioeconomic status in developing countries and lower socioeconomic status in developed countries have been associated with increased diabetes risk.
Whereas type 2 diabetes remains less prevalent in children and adolescents than in any other age group, the trend toward increased obesity and decreased physical activity in youth, especially in industrialized countries, is accompanied by an alarming increase in the incidence and prevalence of pediatric type 2 diabetes, especially among ethnic minorities. For example, the population-based multicenter SEARCH for Diabetes in Youth Study reported that among 1530 youth aged 10 to 19 years with newly diagnosed diabetes, type 2 diabetes accounted for 14.9% of diabetes mellitus cases in non-Hispanic white Americans, 46.1% in Hispanics, 57.8% in African Americans, and 86.2% in Native Americans.
Factors Contributing to the Increasing Incidence and Prevalence of Type 2 Diabetes Mellitus
The global prevalence of overweight (body mass index [BMI] ≥25 kg/m 2 but <30 kg/m 2 ) and obesity (BMI ≥30 kg/m 2 ) continues to increase. The World Health Organization estimates that approximately 1.6 billion adults were overweight and at least 400 million were obese worldwide in 2005 and projects an increase to 2.3 billion overweight and 700 million obese by 2015. The risk of diabetes is proportionally increased with both the severity and duration of obesity, as more people are becoming obese earlier in life and rates of extreme obesity (BMI ≥35 kg/m 2 ) are rising. Beyond BMI, measures of increased abdominal obesity, such as waist circumference and waist-to-hip ratio, are even better predictors of diabetes risk.
It has been difficult to ascertain the role of dietary factors independent of body weight, but the increased availability of foods high in fat, low in fiber, and with a high glycemic load has been associated with increased risk for type 2 diabetes. Increased consumption of red meats in general and processed meats in particular has also been associated with a higher risk of diabetes. Other dietary factors may also play a role. The rise in diabetes prevalence coincides historically with increasing dietary intake of fructose-rich foods; but in spite of data showing that diets very high in fructose induce insulin resistance and diabetes in laboratory animals, whether high fructose intake plays a role in the pathogenesis of diabetes in humans has not been established.
In spite of recent indications that leisure-time physical activity may be stable or even increasing in the United States, overall levels have declined significantly during the past half-century, and the trend toward an increasingly sedentary lifestyle among Americans continues. Physical inactivity has been associated with an increased risk for development of type 2 diabetes even after adjustment for BMI. Conversely, even modest levels of physical activity intensity and duration are associated with decreased risk. For example, among more than 30,000 women (47% of the Nurses’ Health Study cohort) who reported that walking was their sole physical activity, diabetes risk decreased significantly across quintiles of energy expenditure, calculated from walking time and pace. In a high-risk population, the Strong Heart Study of 1651 Native Americans reported an odds ratio for incident diabetes reduced by one third (one fourth after adjustment for BMI) in participants who engaged in even modest amounts of physical activity compared with a control group of no physical activity.
As the incidence and prevalence of type 2 diabetes increase with age, a global trend toward population aging may account for a portion of the increase in diabetes prevalence. Life expectancy at birth in developed countries has increased by approximately 25 years in the twentieth century alone, and even though life expectancy in developing countries has increased at a much slower pace, it is projected that about 80% of the world population older than 60 years will live in developing countries by 2050, with continued aging of the world population projected through the first half of the twenty-first century.
Diagnostic Criteria and Improved Detection
With the thresholds for diagnosis of diabetes becoming more inclusive in recent years and continued improvements in population-based screening, more persons are being diagnosed, with relative increases in the ratio of diagnosed to undiagnosed diabetes.
Type 1 Diabetes Mellitus
Type 1 diabetes results from absolute insulin deficiency due to autoimmune destruction of the insulin-producing beta cells within the islets of the endocrine pancreas. In addition to the obvious glucose abnormalities, patients with type 1 diabetes are also predisposed to ketoacidosis because of unrestrained lipolysis and ketogenesis when insulin falls to undetectable or nearly undetectable levels. Indeed, the diagnosis is commonly made in this setting, often in the context of a superimposed intervening illness that increases counterregulatory stress hormones, such as epinephrine and cortisol.
The diagnosis of type 1 diabetes is straightforward in a patient with hyperglycemia (defined according to the criteria in Table 21-1 ) who is both young and lean, especially one presenting with ketosis or catabolic features, such as weight loss. Serologic confirmation with autoantibodies is rarely necessary. Most patients with type 1 diabetes are, however, seropositive for one or more of the following: anti–glutamic acid decarboxylase antibodies, anti–islet cell antibodies 512, and anti-insulin antibodies. When the diagnosis is less clear, such as in an obese child or in a lean older individual, measurement of these serum markers may prove clinically useful.
Type 1 diabetes can be treated only with insulin, and patients are optimally managed with multiple injections per day, involving both basal and prandial components, or continuous subcutaneous insulin infusion (i.e., an insulin pump). Oral agents are essentially ineffective and have no standard role in the therapy for this disease.
The incidence of type 1 diabetes is low, compared with that of type 2 diabetes, and varies widely between populations, 0.1/100,000 per year in China and Venezuela to approximately 37/100,000 per year in Sardinia and Finland. Although it is typically diagnosed in children, type 1 diabetes may occur at any age; the rapidity of loss of beta cells is inversely proportional to age, with more gradual development in older individuals during many years. The term latent autoimmune diabetes of adulthood has gained favor to describe this form, which initially presents with relatively mild hyperglycemia, often successfully treated with oral agents. Over time, however, insulin deficiency dominates the clinical picture, with labile blood glucose control, insulin dependency, and a propensity for ketosis.
Although it is clearly of autoimmune origin, the underlying cause or causes of type 1 diabetes remain poorly understood. Defined risk factors include certain HLA types, such as DR3 or DR4, and a family history, although the latter is not as potent a factor as it is in type 2 diabetes. For example, children born to men with type 1 diabetes have about a 6% risk for development of the disease. Children born to women with type 1 diabetes have a lower risk, ranging from 1% to 4%, depending on the age of the mother at delivery. Several viruses have also been implicated in the pathogenesis of type 1 diabetes, including mumps, rubella, cytomegalovirus, and coxsackievirus, but they do not appear to be involved in more than a small minority of cases. The possibilities that dietary factors, micronutrient status, and the individual’s intestinal flora may play a role in the predisposition to type 1 diabetes are under active investigation.
Diabetes Mellitus as Risk Factor for Cardiovascular Disease
The World Health Organization estimated that as many as 2.9 million deaths worldwide in the year 2005 could be attributed to diabetes, with more than half of these attributable to CVD. Of the 284,000 deaths attributable to diabetes in the United States in 2007, about two thirds had CVD as the primary cause of death. In spite of a trend for reduced all-cause and cardiovascular mortality in patients with diabetes in the United States during the past half-century, mortality remains approximately twofold higher in those with versus without diabetes ( Fig. 21-3 ).
Diabetic Vascular Disease
Diabetes mellitus is a major risk factor for atherosclerosis, clinically manifested as diabetic macrovascular complications. The risk of ischemic heart disease is twofold to fourfold higher in people with type 2 diabetes compared with those without diabetes, with myocardial infarction (MI) being the number one cause of death. In the Framingham Heart Study between 1976 and 2001, cardiovascular mortality among those with and without diabetes was 6.8 and 2.4 per 1000 person-years, respectively. The risk of MI in patients with diabetes with no history of prior coronary events is, at least in some populations, similar to the risk of MI in patients without diabetes but with prevalent coronary artery disease. For example, in a Finnish population-based study of 1059 patients with type 2 diabetes and 1373 without diabetes aged 45 to 64 years, the 7-year incidence of first MI in those with versus without diabetes was 20.2% versus 3.5%, respectively, and the incidence of recurrent MI was 45% and 18.8%, respectively.
This and other similar evidence led the National Cholesterol Education Program (NCEP) Expert Panel to recommend that type 2 diabetes be managed as a coronary heart disease equivalent for the purpose of low-density lipoprotein cholesterol (LDL-C) control. However, more recent data from clinical trial observations, including studies enrolling patients with newly diagnosed diabetes as well as trials of patients with more advanced type 2 diabetes at high CVD risk, suggest that the CVD risk for diabetes is more intermediate, with risk projected during 10 years ranging between 8% and 20%, contrasted with the “coronary disease equivalent” risk of >20%/10 years.
Type 1 diabetes is also associated with significantly increased cardiovascular risk, especially in younger patients. Data from the Diabetes U.K. Cohort, an observational study of 23,751 patients diagnosed with type 1 diabetes in Great Britain, showed that mortality from ischemic heart disease in men and women older than 40 years with type 1 diabetes is increased 4-fold and 7-fold, respectively, compared with the general population; in those younger than 40 years, the risk increase is 9-fold for men and >40-fold for women. The Pittsburgh Epidemiology of Diabetes Complications Study, a single-center observational study of 906 patients with type 1 diabetes, demonstrated ~15% incidence of CVD after 30 years in this relatively young cohort (median age at onset, 8.5 years).
Having type 2 diabetes doubles one’s risk of stroke, even after adjustments for other risk factors (hypertension, dyslipidemia), and increases by 15-fold the risk of lower extremity amputation due to peripheral arterial disease. These risks have also been demonstrated in type 1 diabetes, the earlier onset of which results in significantly higher relative risk in younger age groups. For example, the risk of cerebrovascular mortality in the Diabetes U.K. Cohort was increased by approximately twofold in patients with type 1 diabetes 60 years and older compared with participants without diabetes but by more than fivefold and sevenfold, respectively, in men and women aged 20 to 39 years. Finally, atherosclerotic events such as MI and stroke are associated with higher short- and long-term mortality, higher rate of recurrence, and worse overall prognosis in the context of diabetes.
Approximately 70% of patients with type 2 diabetes have hypertension (more than double the prevalence in the general population), which further increases their CVD risk. For example, an observational analysis of 3642 patients with type 2 diabetes enrolled in the United Kingdom Prospective Diabetes Study (UKPDS) demonstrated a positive adjusted correlation between mean systolic blood pressure and the risk of MI, stroke, and heart failure ( Fig. 21-4 ). Hypertension is also an independent risk factor for chronic kidney disease, which in turn may exacerbate CVD risk, resulting in a vicious circle.
People with type 2 diabetes have a twofold to fivefold increased risk of congestive heart failure (CHF) compared with those without diabetes and have worse outcomes once CHF has developed. Diabetes increases the incidence of CHF following the entire spectrum of acute coronary syndromes and remains an independent predictor of CHF even after adjustment for the increased prevalence of ischemic heart disease among patients with diabetes. The increased CHF observed in diabetes is multifactorial and includes more prevalent systolic and diastolic dysfunction due to both circulatory impairments and derangements of myocardial metabolism (diabetic cardiomyopathy).
Sex and Ethnic Differences
Available data suggest that relative cardiovascular risk is higher in women compared with men with type 1 or type 2 diabetes and that diabetes reduces the sex differences in CVD incidence that otherwise exist in the nondiabetic population. For example, type 2 diabetes in the Framingham Heart Study increased the risk of CVD mortality by 3.5- to 5-fold in women and 2- to 3-fold in men.
The risk for diabetic complications is also not uniform across ethnic groups. Data from an observational study of 62,432 diabetic patients showed that African Americans, Asians, and Hispanics have a relatively lower risk of MI than that of whites (adjusted hazard ratio [HR], 0.56, 0.68, and 0.68, respectively). Asians and Hispanics have a lower risk of stroke and CHF than that of both whites and African Americans (HR, 0.76 and 0.72, stroke; 0.70 and 0.61, CHF). In Pima Indians, despite a very high overall prevalence of diabetes, the incidence of fatal coronary heart disease is comparatively low.
Potential Mechanistic Links Between Diabetes Mellitus and Cardiovascular Disease
Diabetic Macrovascular Disease
The complex mechanistic interrelationships between type 2 diabetes and atherosclerosis have been the focus of extensive and ongoing investigation. A detailed review of this field is beyond the scope of this chapter, but a number of published review articles provide excellent summaries of the present knowledge in this area. Among the principal vascular perturbations associated with type 2 diabetes is increased endothelial dysfunction, driven largely by dysregulated nitric oxide biology, indirect and direct vascular effects of advanced glycation end products, direct adverse effects of increased circulating nonesterified free fatty acids, and increased systemic inflammation and aberrant leukocyte-endothelial interactions, among others.
Compounding the direct vascular effects of diabetes are a number of perturbations in the proteo-fibrinolytic system and platelet biology yielding a constitutive prothrombotic milieu. These abnormalities include increased circulating tissue factor, factor VII, von Willebrand factor, and plasminogen activator inhibitor 1, with decreased levels of antithrombin III and protein C. In addition, disturbances of platelet activation, aggregation, morphology, and life span have been well described, further contributing to increased thrombotic potential as well as acceleration of atherosclerosis.
Finally, diabetic dyslipidemia is a major contributor to the pathogenesis and progression of atherosclerosis. Diabetic dyslipidemia is a constellation of metabolically interconnected lipid and lipoprotein abnormalities, including increased plasma triglycerides, decreased high-density lipoprotein (HDL), and a modest increase in low-density lipoproteins (LDL), with larger proportions of atherogenic LDL such as small dense LDL and oxidized LDL. Decreased HDL impairs reverse cholesterol transport (the movement of cholesterol from peripheral tissues to the liver) and reduces the anti-inflammatory and antioxidant effects of HDL in the circulation. Elevated levels of apolipoprotein B–containing lipoproteins, such as LDL and triglyceride-rich very-low-density lipoproteins (VLDL), and increased remnant lipoproteins (produced by the hydrolysis of VLDL and chylomicrons) directly promote atherosclerosis.
Several pathophysiologic processes, individually or in combination, have been proposed to contribute to the pathogenesis of diabetic cardiomyopathy and CHF. Principal among these processes are abnormal insulin action at the level of the cardiac myocyte coupled with increased circulating free fatty acids, resulting in aberrant myocardial metabolism with accumulation of free fatty acids and triglycerides, and the generation of reactive oxygen species and toxic lipid metabolites, termed cardiac lipotoxicity.
In addition, diabetes is associated with impaired myocellular metabolic substrate switching, deleteriously perturbing the balance between free fatty acid and glucose metabolism under periods of stress such as ischemia, with increased free fatty acid metabolism increasing myocardial oxygen consumption. Hyperglycemia may exacerbate these effects by inhibiting free fatty acid oxidation, with excess intracellular glucose resulting in nonenzymatic protein glycation, formation of intracellular and extracellular advanced glycation end products, and resultant adverse mechanical and metabolic consequences.
Prevention of Type 2 Diabetes Mellitus
The first line of defense against diabetic CVD is prevention of diabetes. The identification of modifiable risk factors and populations at increased risk for development of type 2 diabetes has made diabetes prevention or delay theoretically feasible. Preventive measures should specifically target people at high risk (see Table 21-2 ) but could also be applied to the general population.
A combined intervention consisting of diet, physical activity, and weight loss reduces diabetes risk, as summarized in Figure 21-5 . The Finnish Diabetes Prevention Study included 522 overweight subjects with impaired glucose tolerance randomized to either usual care or intensive lifestyle intervention (aimed at reduction of total intake of fat to <30% and saturated fat to <10% of energy consumed, increase in the intake of fiber to >15 g per 1000 kcal, moderate exercise for at least 30 minutes per day, and at least 5% weight reduction). After a median follow-up of 4 years, the relative risk of diabetes was reduced by 58% in the intervention group, with sustained benefit observed up to 3 years after the study.
The U.S. Diabetes Prevention Program randomized 3234 people with impaired glucose tolerance or impaired fasting glucose to placebo control, metformin, or an intensive lifestyle modification program aimed at a minimum of 150 minutes of exercise per week, low-fat diet, and 7% weight loss. The incidence of diabetes was reduced by 58% in the lifestyle modification group and by 31% in the metformin group after an average follow-up of 2.8 years.
In other high-risk populations, the Indian Diabetes Prevention Program reported a 28.5% reduction in progression to diabetes with lifestyle modification (diet and exercise) for 3 years in Asian Indians with impaired glucose tolerance, and a randomized trial of Japanese men with impaired glucose tolerance observed for 4 years reported a 67.4% reduction in the risk of diabetes with diet, exercise, and sustained weight loss. The cluster-randomized (by clinic rather than by participant) Chinese Da Qing study found that at-risk individuals in the combined intervention groups (diet, exercise, and diet plus exercise) had a 51% lower incidence of diabetes compared with the control group during 6 years of active intervention and a 43% lower incidence during a 20-year period including 14 years of post-trial follow-up. Despite the variability in risk reduction across these studies, each has demonstrated that a combination of diet, physical activity, and weight loss significantly reduces or delays the development of type 2 diabetes, with several showing sustained effects. On the basis of the available evidence, the ADA recommends reduced intake of dietary calories and fat, regular moderate-intensity physical activity (≥150 min/week), and sustained weight loss (≥7% of body weight) for diabetes prevention in people at risk.
Independent of weight effects, dietary composition also plays a role in type 2 diabetes risk. Overall, diets high in red and processed meats and saturated fats have been associated with increased risk for type 2 diabetes, and conversely, low–saturated fat, high-fiber diets containing whole grains, fruits, vegetables, and fish have been associated with reduced risk of type 2 diabetes. The dietary composition recommended by the ADA for diabetes prevention includes reduced intake of fat, increased dietary fiber (14 g fiber per 1000 kcal), and whole grains (at least half of grain intake), and foods with low glycemic index (generally less processed) are encouraged.
Physical Activity and Exercise
Even though physical activity (overall) and exercise (planned and structured leisure-time physical activity) may have only a modest effect on weight if they are not accompanied by dietary measures, increasing physical activity, which directly improves insulin sensitivity in skeletal muscle, may be beneficial for diabetes prevention independent of weight loss. Prospective cohort studies of 21,271 American men aged 40 to 84 years participating in the Physicians’ Health Study and of 7735 British men aged 40 to 59 years, all free of diabetes at baseline, showed that individuals who engaged in moderate levels of physical activity had a significantly lower risk for diabetes compared with those who were physically inactive, even after adjustment for age and BMI.
Similarly, data from the Nurses’ Health Study (70,102 nondiabetic women aged 40 to 65 years observed for 8 years) showed that higher quintiles of physical activity were associated with decreased diabetes risk, and the trend remained significant after adjustment for BMI.
On the basis of the available evidence, regular physical activity (at least 30 min/day on most days or 150 min/week as recommended by the ADA) is important for diabetes prevention. It should be recommended to individuals at risk in addition to other weight loss strategies.
Treatment of Obesity
When lifestyle modification alone fails to reduce body weight in obese patients, pharmacologic or surgical intervention may be considered. Orlistat is a gastric and pancreatic lipase inhibitor that reduces fat absorption by ~30%. In a 4-year trial of 3305 obese patients randomized to lifestyle changes plus either orlistat or placebo, orlistat was associated with a greater weight loss (5.8 kg versus 3.0 kg) and a lower incidence of diabetes (9.0% versus 6.2%; 37.3% relative risk reduction).
Bariatric surgery, which modifies the gastrointestinal tract anatomy through a variety of techniques to reduce calorie intake and to affect weight loss, has also been shown to have myriad beneficial effects on measures of glucose and lipid metabolism, including reducing the risk of diabetes. In addition to weight reduction, bariatric procedures are associated with beneficial effects on the secretion of nutrient-responsive gut hormones, including incretins, ghrelin, and peptide YY, that may improve insulin responsiveness independent of weight loss.
Several studies support the use of bariatric surgery as an effective strategy for weight loss in severely obese patients. The prospective controlled nonrandomized Swedish Obese Subjects trial reported mean weight reductions of 25% for the Roux-en-Y procedure and 13% to 16% for the restrictive procedures at 10 years after intervention, compared with a trend for further weight gain in the control group (nonsurgical, nonstandardized obesity management). The incidence of diabetes at 2 and 10 years was 1% and 7%, respectively, in the overall bariatric surgery group, compared with 8% and 24%, respectively, in the control group, with suggestions of modest long-term mortality benefit. However, the potential benefits of bariatric surgery should be carefully weighed against its significant risk for perioperative complications (including early death) and long-term gastrointestinal and nutritional adverse effects.
Drug Therapy for Prevention of Diabetes Mellitus
The most compelling evidence for pharmacologic prevention of diabetes comes from a study of 2155 individuals with impaired glucose tolerance randomized to either metformin or placebo in the Diabetes Prevention Program. After a mean follow-up of 2.8 years, the incidence of diabetes was 4.8% in the metformin group versus 7.8% in the placebo group, representing a 31% relative risk reduction. Notably, however, metformin was indistinguishable from placebo in participants older than 60 years as well as in those whose fasting glucose concentration was <110 mg/dL and in those whose BMI was <35 kg/m 2 .
To determine whether the effect of metformin was transient (“masking” of diabetes) or sustained (true prevention), 1274 participants who had not developed diabetes by the end of the Diabetes Prevention Program were subjected to an oral glucose tolerance test 1 to 2 weeks after the discontinuation of study medication. There was an increased incidence of diabetes in the metformin group during this washout period (indicating that masking did occur in those cases), but when the washout data were included in the overall analysis, the diabetes risk reduction with metformin remained significant at 25%.
On the basis of these observations, metformin is the only drug recommended by the ADA to be considered for diabetes prevention. It is specifically recommended for high-risk patients with both impaired fasting glucose and impaired glucose tolerance who are younger than 60 years, have a BMI >35 kg/m 2 , and have other diabetes risk factors (HbA1c >6.0%, hypertension, low HDL-C, high triglycerides, family history).
The Study To Prevent Non–Insulin-Dependent Diabetes Mellitus (STOP-NIDDM) randomized 1429 participants with impaired glucose tolerance to either acarbose or placebo. After a mean follow-up of 39 months, the cumulative incidence of diabetes was 32% in the acarbose group and 42% in the placebo group, representing a 25% relative risk reduction. However, the incidence of diabetes during 3 months of single-blind post-trial follow-up on placebo was 45% higher in the former acarbose group, suggesting a significant masking component.
Another drug in this class, voglibose, was studied in a randomized trial of 1780 Japanese subjects at high risk for diabetes and was found to reduce the incidence of diabetes by 40% after a short mean follow-up of 48 weeks. This trial was stopped early because of apparent benefit, which is controversial and could in fact overestimate the effects of the active medication. Of note, agents in this class have significant gastrointestinal side effects that limit their broad appeal.
By virtue of their insulin-sensitizing effects through activation of the nuclear receptor peroxisome proliferator-activated receptor γ (PPARγ) that regulates gene transcription, the thiazolidinediones (TZDs), including rosiglitazone and pioglitazone, target the principal pathologic underpinning of type 2 diabetes. In addition, over longer duration, the achieved insulin sensitization may improve beta-cell preservation, which could plausibly enhance their effects on prevention or delay of progression to diabetes. Thus, the TZDs appear promising in this regard, with supportive data deriving from a number of randomized controlled clinical trials enrolling patients at exaggerated diabetes risk, including Hispanic women with gestational diabetes, and large trials of subjects with impaired fasting glucose or impaired glucose tolerance ( Fig. 21-6 ). The magnitude of the treatment effects in these trials was large, ranging from 55% to 81% relative risk reduction for diabetes, but commensurate with the risk reduction achieved by lifestyle intervention described before.
Despite the robust effects observed across the TZD drug class and across trials, however, neither of the two currently available TZDs has a product label indication for this purpose, and they are not recommended for diabetes prevention in society guidelines. The absence of recommendation stems from a number of considerations, including ongoing concerns about adverse effects such as increased heart failure and fracture risk as well as the current expense of these drugs. In addition, data are discordant regarding the effects of these drugs on atherosclerotic disease risk, and none of the diabetes prevention trials was designed or powered to assess effects on macrovascular or microvascular disease endpoints. On the basis of these considerations, TZDs are not presently recommended for prevention of type 2 diabetes.
Management of Hyperglycemia to Prevent Cardiovascular Disease
Glucose and Cardiovascular Disease Risk: Epidemiology
Extensive epidemiologic evidence has linked measures of hyperglycemia, including fasting plasma glucose, random plasma glucose, 2-hour post-challenge plasma glucose, and HbA1c, with increased overall and CVD mortality, with associations extending well into what is considered the normal range for glucose. One large meta-analysis by Levitan and coworkers of 38 investigations found that overall, those with the highest measures of glucose experience approximately one third more CVD events than those with the lowest (RR, 1.36; 95% CI, 1.23-1.52). This relationship persisted even when diabetic subjects were excluded from the analysis (RR, 1.26; 95% CI, 1.11-1.43) ( Fig. 21-7 ). Adjustment for CVD risk factors attenuated the relationship to some degree (RR, 1.19; 95% CI, 1.07-1.32). A second and separate meta-analysis by Selvin and colleagues of 13 observational studies of diabetic patients (N = 9123) found a pooled relative risk for CVD of 1.18 in type 2 diabetes (95% CI, 1.10-1.26) and 1.15 in type 1 diabetes (95% CI, 0.92-1.43) for each 1% increase in HbA1c.
In a more recent analysis using NHANES III data from 19,025 adult participants (baseline survey in 1988-1994 with follow-up through 2000), Saydah and coworkers found that higher levels of HbA1c were associated with increased mortality, including that due to heart disease. After adjustment for other CVD risk factors, the hazard ratio for adults with HbA1c ≥8% versus <6% was 2.59 (95% CI, 1.88-3.56) and 3.38 (95% CI, 1.98-5.77) for all-cause and cardiovascular mortality, respectively. The comparative hazard ratios for adults with a diagnosis of diabetes were 1.68 (1.03-2.74) and 2.48 (1.09-5.64), respectively. However, in this analysis, among those without diagnosed diabetes, no significant association between either all-cause or cardiovascular mortality and HbA1c category was found.
Within the context of a clinical trial, the largest epidemiologic analysis comes from the UKPDS, involving 5102 patients with newly diagnosed type 2 diabetes. In line with the estimates of Selvin and colleagues, Stratton and collaborators have shown that each 1% reduction in HbA1c was associated with a 14% reduction in MI events. Of note, however, the graded association between HbA1c and microvascular endpoints was far steeper than that for MI ( Fig. 21-8 ).
Overview of Current Antihyperglycemic Drug Classes
A variety of both oral and injectable antihyperglycemic agents are used for blood glucose control in patients with type 2 diabetes. In general, these include drugs that stimulate insulin release, improve the body’s response to insulin, or delay carbohydrate absorption. In those patients not responding adequately to these agents, typically used in combination, insulin therapy may be used, although patients usually do not require the highly intensive programs needed in those with type 1 diabetes.
There are now 11 individual drug categories approved for glucose lowering in patients with type 2 diabetes in the United States. Their mechanisms of action, potency, advantages, and disadvantages are summarized in Table 21-3 .
|Drug Class||Examples||Underlying Mechanism||Main Metabolic Effects||↓A1c||Advantages||Disadvantages|
|Closes K ATP channels||↑ Pancreatic insulin secretion||∼1%-2%||Microvascular risk||Hypoglycemia |
Ischemic preconditioning (?)
|Closes K ATP channels||↑ Pancreatic insulin secretion||∼1%-1.5%||More physiologic than sulfonylureas |
↓ Postprandial glucose
Ischemic preconditioning (?)
|Biguanides||Metformin||Activates AMP kinase||↓ Hepatic glucose production||∼1%-2%||No hypoglycemia |
↓ CVD events
|Gastrointestinal side effects (diarrhea) |
Multiple contraindications to consider
|Activates PPARγ||↑ Peripheral insulin sensitivity||∼1%-1.5%||No hypoglycemia |
↓ CVD events (?) (pio)
↓ Blood pressure
|Weight gain |
Edema, heart failure
Bone fractures (women)
↑ LDL-C (↑ particle size)
Rosiglitazone controversy in coronary heart disease
|α-Glucosidase Inhibitors||Acarbose |
|Blocks small intestine α-glucosidase||↓ Intestinal carbohydrate absorption||∼0.5%-1%||No hypoglycemia |
↓ Postprandial glucose
↓ CVD events (?)
|Gastrointestinal side effects (flatulence) |
|Glucagon-like peptide 1 (GLP-1) agonists||Exenatide||Activates GLP-1 receptors||↑ Pancreatic insulin secretion |
↓ Pancreatic glucagon secretion
Delays gastric emptying
|∼1%||Weight loss |
↓ Postprandial glucose
Cardiovascular benefits (?)
|Gastrointestinal side effects (nausea, vomiting) |
|Amylinomimetics||Pramlintide||Activates amylin receptors||↓ Pancreatic glucagon secretion |
Delays gastric emptying
|∼0.5%||Weight loss |
↓ Postprandial glucose
|Gastrointestinal side effects (nausea, vomiting) |
|Dipeptidyl peptidase (DPP) 4 inhibitors||Sitagliptin |
|Inhibits DPP-4 |
↑ Endogenous incretin levels
|↑ Pancreatic insulin secretion |
↓ Pancreatic glucagon secretion
|∼0.6%-0.8%||No hypoglycemia||Urticaria, angioedema, pancreatitis (?)|
|Bile acid sequestrants||Colesevelam||Binds bile acid cholesterol||?||∼0.5%||No hypoglycemia |
|Gastrointestinal side effects (constipation) |
|Dopaminergic receptor 2 (D 2 ) agonists||Bromocriptine||Activates dopaminergic receptors||Modulates hypothalamic circadian organization |
↓ Hepatic glucose production
|∼0.5%||No hypoglycemia||Gastrointestinal side effects (nausea) |
|Insulin||Human NPH |
Premixed (various types)
|Activates insulin receptors||↑ Peripheral glucose disposal |
↓ Hepatic glucose production
|No limit||↓ Microvascular risk |
The sulfonylurea drug class is the oldest, in use since the 1950s. Their main side effect is hypoglycemia, which activates an adrenergic response and therefore can be potentially deleterious to the cardiovascular system. Direct negative effects on the heart from these agents have also been proposed because the receptor they activate, the sulfonylurea receptor, is expressed in cardiomyocytes, and activation causes closure of ATP-dependent potassium channels (K ATP ). Because K ATP closure in myocytes during ischemia may impair preconditioning, sulfonylureas may theoretically exacerbate cardiac injury. Although several animal models have suggested such an effect, the degree to which the drugs bind to K ATP channels in the heart is much less than in pancreatic beta cells, and there are no convincing human data to suggest harm. Sulfonylureas are therefore widely considered to be safe in diabetic patients, including those with coronary artery disease, although this remains consensus opinion in the absence of rigorous data from clinical outcomes trials. Logically, however, these drugs should not be used in the setting of acute coronary syndromes.
Metformin, a biguanide, reduces glucose primarily by attenuating hepatic glucose output, probably by activation of the enzyme AMP kinase. In the UKPDS randomized trial, the risk of MI was reduced by 39% ( P = 0.01) with metformin compared with those receiving conventional care (diet alone). Such an effect was not convincingly demonstrated in the group assigned to sulfonylureas or insulin (relative risk reduction, 16%; P = 0.052). On the basis of these data, paired with a low hypoglycemia risk, modest weight loss, high tolerability with few adverse effects, and low cost, metformin is widely favored as initial monotherapy in most patients with type 2 diabetes. Whereas the use of metformin is cautioned for patients with decompensated heart failure and contraindicated in patients with advanced renal disease because of concern for lactic acidosis, the aggregate data suggest that this risk is quite small or possibly even nonexistent.
The TZDs (pioglitazone, rosiglitazone) activate PPARγ and improve insulin sensitivity by increasing glucose uptake by peripheral tissues, mainly skeletal muscle. They also augment lipogenesis by inducing differentiation of preadipocytes. Their use is associated with a glucose-lowering effectiveness on the same order as that seen with sulfonylureas and metformin and is not associated with hypoglycemia. Pioglitazone, but probably not rosiglitazone, also has beneficial lipid effects, including a 10% to 15% increase in HDL-C and a commensurate decrease in triglycerides. Unfortunately, TZDs are associated with weight gain and fluid retention as well as increased fracture risk in women. The fluid retention is associated with incident or worsening heart failure. The incidence of peripheral edema in TZD-treated patients is approximately 5% to 10%, more so when they are used in conjunction with insulin. The risk of heart failure is approximately 1%, also higher in insulin-treated patients. The relative risk of heart failure in patients taking TZDs is approximately 1.7 compared with a non-TZD regimen, with higher reported risk increase with rosiglitazone than with pioglitazone.
There is little cardiovascular information available for the other antihyperglycemic therapies presently marketed for diabetes. There are very few data with the newer incretin-based therapies, such as the glucagon-like peptide 1 (GLP-1) agonists and the dipeptidyl peptidase 4 inhibitors. There is great interest in the potential cardiovascular benefit of GLP-1 agonists (e.g., exenatide) because these agents result in substantial weight loss in some patients, with the early demonstration of associated improvement in various cardiovascular risk factors. Studies suggesting a benefit on ventricular function in heart failure and after acute coronary events are very preliminary.
Of note, regulatory agencies around the globe have recently begun requiring clinical trial assessment of the CVD effects of diabetes drugs under development, with a requirement to demonstrate a nominal degree of cardiovascular safety before drug approval.
Type 2 Diabetes Mellitus
The UKPDS demonstrated microvascular risk reduction from more intensive glycemic control (relative risk reduction, 21% to 34%) in 5102 patients with newly diagnosed type 2 diabetes. Patients were randomized to treatment with sulfonylurea or insulin as first-line therapy compared with diet alone. During a mean follow-up of 10 years, mean HbA1c was 7.0% in actively treated patients versus 7.9% in the control group, but the effect on macrovascular outcomes (nonfatal and fatal MI and sudden death) did not reach statistical significance (16% risk reduction; P = 0.052). In a randomized substudy (N = 753) of the UKPDS, initial monotherapy in overweight patients (>120% of ideal body weight) with metformin resulted in improved macrovascular outcomes, an outcome not observed within the main trial (see earlier).
From the UKPDS came the original notion that glucose control, although apparently important for microangiopathy, had more obscure effect on CVD complications. Given the metformin findings, it also raised concerns about whether the method by which glucose is lowered may play some role in altering this disease’s cardiovascular risk equation.
Long-term post-trial follow-up of the UKPDS cohort has demonstrated cardiovascular benefit from previous intensive glucose control. Despite a loss in the HbA1c differences between the groups, during 10 years of post-trial monitoring, the patients previously intensively treated with sulfonylurea or insulin experienced relative risk reductions for any diabetes-related endpoint (9%; P = 0.04) and microvascular disease (24%; P = 0.001), whereas decreases in the relative risk for MI (15%; P = 0.01) and all-cause mortality (21%; P = 0.01) eventually emerged. In the overweight metformin-treated patients, corresponding reductions in any diabetes-related endpoint (21%; P = 0.01), MI (33%; P = 0.005), and all-cause mortality (27%; P = 0.002) were observed. These important results demonstrate a “glycemic memory” concept whereby the effects of intensive treatment continued beyond the randomization phase of the study, eventually resulting in beneficial macrovascular outcome differences.
ACCORD, ADVANCE, VADT
More recently, three randomized clinical trials (ACCORD, ADVANCE, and VADT ) formally tested the cardiovascular effect of more versus less intensive glucose control (HbA1c target of 6% to 6.5% versus 7% to 9.0%) during a period of 3 to 5 years, all three failing to demonstrate significant benefit ( Table 21-4 ). In the ACCORD trial, cardiovascular mortality was actually increased in the intensive group, although the explanation for this phenomenon remains enigmatic. Hypoglycemia was more common in patients who died in both intensive and standard control groups, but post hoc analyses have not been able to demonstrate a causative relationship. Notably, more than three of four intensively treated patients received insulin during the course of the trial, and the majority of patients were prescribed at least three oral agents simultaneously. Whether such polypharmacy may have played a role in the study results is also not clear.
|Age (mean, years)||62||66||60|
|BMI (mean, kg/m 2 )||32||28||31|
|Follow-up (mean, years)||3.5||5||5.6|
|HbA1c target||<6.0% vs. 7.0%-7.9%||≤6.5% vs. “standard”||<6% vs. 8%-9%|
|Baseline HbA1c (mean)||8.3%||7.5%||9.4%|
|Endpoint HbA1c (mean)||Intensive |
|Severe hypoglycemic events||Intensive |
|Weight change||Intensive |
|Major macrovascular or microvascular event||Not reported||0.9 (0.82-0.98); P = 0.01||0.88 (0.74-1.05); P = 0.14|
|Nonfatal MI or stroke, cardiovascular death||HR, 0.9 (0.78-1.04); P = 0.16||0.94 (0.84-1.06); P = 0.32||Not reported|
|All-cause mortality||HR, 1.22 (1.01-1.46); P = 0.04||0.93 (0.83-1.06); P = 0.28||1.07 (0.81-1.42); P = 0.62|
|Nonfatal MI||HR, 0.76 (0.62-0.92); P = 0.004||0.98 (0.77-1.22); P = NS||0.82 (0.59-1.14); P = 0.24|