Cardiovascular Disease in Diabetes
The majority (65% to 80%) of people with diabetes will die from heart disease (1,2). Compared to people without diabetes, people with diabetes (especially women) are at higher risk of developing heart disease, and at an earlier age. A high proportion of deaths occur in patients with no prior signs or symptoms of cardiovascular disease (CVD). Furthermore, people with diabetes have a high prevalence of silent myocardial ischemia, and almost one-third of myocardial infarctions (MIs) occur without recognized or typical symptoms (silent MIs) (3). The goals of screening are to improve life expectancy and quality of life by preventing MI and heart failure through the early detection of coronary artery disease (CAD).
Role of Stress Testing
Exercise stress testing is useful in patients at high risk of CAD for the assessment of prognosis and the identification of individuals who may benefit from coronary artery revascularization to improve long-term survival. The most predictive clinical observation for CAD in the person with or without diabetes is a history of chest pain or discomfort, but these features will be absent in a significant number (20% to 50%) of people with diabetes (4–10). Clinical findings, such as dyspnea on exertion, resting electrocardiogram (ECG) abnormalities or multiple risk factors for atherosclerosis, may also indicate the presence of CAD. Recognition of such features is of clinical importance, as the outcome of CAD events is worse in people with diabetes when shortness of breath is the primary symptom (4).
The presence of CAD risk factors and resting ECG abnormalities identify patients with diabetes at increased risk of important CAD burden and abnormal stress ECG or perfusion imaging results (11). A resting ECG at the time of diagnosis of diabetes also provides a baseline to which future ECGs can be compared. In patients considered to be at high risk for CAD, a repeat resting ECG may detect changes that result from silent MI and lead to earlier detection of critical CAD. There is evidence that early screening and intervention in people with diabetes and with silent ischemia is beneficial and may improve long-term survival (7,12). Screening with exercise ECG stress testing will find 3-vessel CAD in 13% to 15% of those with abnormal stress test findings (13) and lead to angiography with revascularization in 1% to 3% of asymptomatic individuals (13–15). The Definition of Ischemia in Asymptomatic Diabetes (DIAD) study was prospectively investigating the value of routine adenosine stress myocardial perfusion scanning in asymptomatic patients with type 2 diabetes ≥55 years for the prevention of coronary events (10). The baseline study showed either perfusion defects or stress-induced ECG abnormalities in 22% of patients and large defects in 6%. In this study, multiple risk factors for CAD did not help to predict the patients with positive screening tests for CAD. A substantial portion of the DIAD population was defined as having intermediate/high baseline cardiovascular risk. Nevertheless, their annual cardiac event rate was low and not altered by routine screening for inducible ischemia. Yet, a randomized pilot study on the impact of stress testing to screen for CAD in asymptomatic subjects with diabetes suggested a significant reduction in cardiac death and MI (16). Larger and adequately powered studies are necessary to support this provocative observation before clinical practice is changed. However, it is important to keep in mind that the goals of screening for CAD are to improve life expectancy and quality of life by preventing MI and heart failure through early detection.
The choice of initial stress test should be based on evaluation of the resting ECG, the individual's ability to exercise, and local expertise and technology. There are data with newer technology, but the add-on effect of the latter on prognosis and quality of life is not clear. ECG abnormalities that limit the diagnostic accuracy of a stress ECG include resting ST depression (≥1 mm), left bundle branch block or right bundle branch block, an intraventricular conduction defect with QRS duration >120 ms, ventricular paced rhythm or preexcitation. Individuals with these resting ECG findings should have a stress test with an imaging modality, such as scintigraphic myocardial perfusion imaging or echocardiography. The role of other imaging modalities (anatomical imaging), such as coronary computed tomography (CT), calcium score, etc., in comparison to functional imaging, needs to be determined in individuals with diabetes.
The strongest and most consistent prognostic marker identified during exercise ECG stress testing is the person's maximum exercise capacity (4). Although exercise capacity is decreased in individuals with diabetes (17,18) , it is still of prognostic importance (4). Silent ischemia is most likely to occur in individuals with diabetes who are older (mean age 65 years) and have elevated total cholesterol and proteinuria (14). An ECG with ST-T abnormalities at rest has been shown to be most predictive for silent ischemia (Odds Ratio (OR) 9.27, 95% confidence interval [CI], 4.44-19.38) and was the only significant predictor of silent ischemia in women (14). The relevance of ST-T abnormalities as a predictive factor for silent ischemia emphasizes the importance of recording a resting ECG in most individuals with type 2 diabetes. An abnormal ECG may indicate the need for further investigations and result in the earlier detection and treatment of CAD (14). An abnormal exercise ECG is associated with an annual CAD event rate of 2.1%, compared with 0.97% in subjects with normal exercise ECG (16). Myocardial ischemia (whether silent or symptomatic) detected during exercise stress testing in individuals with diabetes is associated with poorer long-term survival compared to individuals without diabetes (7). Silent MI is common (40%) in older asymptomatic people with type 2 diabetes, but is more frequent (65%) in those with diabetes who also have microalbuminuria (19). People with diabetes and silent ischemia have an annual event rate for CAD of 6.2% (50% of events were new-onset angina and 50% were cardiac death or MIs) (20). Thus, silent MI is a prelude not only to symptomatic ischemia, but also to potentially fatal events. Also, it has been shown in a randomized trial in patients with silent ischemia (the vast majority of whom did not have diabetes) that long-term anti-ischemic drug therapy (∼11 years follow-up) reduces cardiac events (cardiac death, non-fatal MI, acute coronary syndrome, or revascularization) with preservation of ejection fraction (21).
Exercise capacity is frequently impaired in people with diabetes due to the high prevalence of obesity, sedentary lifestyle, peripheral neuropathy (both sensory and motor), and vascular disease in this population. Individuals who cannot adequately exercise on a stress test have a poorer prognosis than those who can, regardless of the reason for this incapacity. Perfusion imaging also provides important prognostic information. Myocardial perfusion imaging has similar predictive value for cardiac death and non-fatal MI in individuals with diabetes as in those without diabetes (22). For those unable to perform an exercise ECG stress test, pharmacologic stress imaging, using dipyridamole, adenosine, or dobutamine testing, is required. Stress echocardiography and stress nuclear imaging have similar values for cardiac events in the general population (23) , but no comparative data are available for the person with diabetes. In a meta-analysis of perfusion imaging, an abnormal scan was predictive of future CAD events in subjects with and without diabetes. However, the cardiac event rate in individuals with diabetes was significantly greater than in those without diabetes (23). The choice of the optimal imaging modality to detect stress-induced MI is best determined by local availability and expertise. The utility of newer CAD diagnostic modalities, such as CT angiography, coronary artery calcium scoring, and cardiac magnetic resonance imaging, is currently unknown in terms of guiding management decisions in patients with type 2 diabetes (24).
CVD in Type 1 Diabetes
Incidence and prevalence of CVD
CVD complications are important causes of morbidity and mortality among individuals with type 1 diabetes which may have been under-recognized in the past. Reported prevalence rates of CVD in type 1 diabetes vary between 3% and 12.4% (25–27). It is important to emphasize that the cardiovascular risk burden and profile of patients with type 1 diabetes differs from type 2 diabetes. The Diabetes UK longitudinal cohort study, including more than 7,000 patients with type 1 diabetes, reported that type 1 diabetes is associated with markedly increased adjusted hazard ratio for major CAD events (median follow-up of 4.7 years) in both men (HR 3.6) and women (HR 9.6). Of such, these risk increments are comparable to those observed in patients with type 2 diabetes (27). Major CVD events occurred in type 1 diabetes on average 10 to 15 years earlier compared with matched nondiabetic controls. Despite the much younger age of onset, the age-adjusted relative risk for CVD in type 1 diabetes is ten times that of the general population (28–30). The Pittsburgh Epidemiology of Diabetes Complications (EDC) study demonstrated that the incidence of major CVD events in young adults with type 1 diabetes (age 28 to 38 years) was 0.98% per year (31) and was as high as 3% per year after age 55 years, making it the leading cause of death in that population (26,27,32). Gender and race/ethnicity are important features of increased risk of CVD; male gender and African-Americans have higher rates of CVD compared to Europeans (31).
Difference from type 2 diabetes
CVD in type 1 diabetes differs from type 2 diabetes, not only in that it presents at a younger age but also in relation to sex, silent presentation and disease severity (28,29). The risk of CAD mortality is comparable in women and men with type 1 diabetes, and there is a high prevalence of silent CAD in young adults with type 1 diabetes, which may be related to cardiac autonomic neuropathy. Finally, the disease process seems to be more severe in type 1 diabetes. Compared with nondiabetic controls, patients with type 1 diabetes are more likely to have severe coronary stenosis, involvement of all 3 major coronary arteries and distal segment disease, resulting in major cardiovascular events with poor outcome and/or early development of heart failure (28,29).
Coronary artery disease and cerebrovascular disease
CAD appears more common than stroke. The cumulative incidence of CAD ranges between 2.1% (26) and 19% (33) depending on the characteristics of the population studied. For the most part, studies report incidences around 15% (27,34,35). Mortality rates from CAD are reported between 6 and 8% (33,35) , are likely higher in men than women (36) , and in those >40 years of age compared to those <40 years of age (36). Stroke is still an important outcome in type 1 diabetes; the cumulative incidence of stroke was 3.3% over 6 years among African-Americans (27) , 5.9% over 20 years in the WESDR (Wisconsin Epidemiologic Study of Diabetic Retinopathy) (34) , and 0.74% per year in the EURODIAB Study (26). Also, prevalence of silent brain infarcts or leukoaraiosis is extremely high (34.5%) in type 1 diabetes (37).
Peripheral vascular disease
Peripheral vascular disease (PVD) is an important vascular complication of type 1 diabetes. Incidence rates of lower extremity amputation vary by age from 3.6 per 1000 person-years among individuals 25 to 44 years of age to as high as 7.2% (38). By age 65, the cumulative probability of PVD is 11% in women and 20.7% in men (39). Compared to the general population, the rate of PVD among those with type 1 diabetes may be very high (39). If one considers ankle-brachial index (ABI) <0.9 as the criterion for the presence of peripheral atherosclerotic disease instead of overt clinical events, 45.6% of participants from the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) study developed PVD (41). Predictors of PVD include increasing age, male gender, history of sores or ulcers, diastolic blood pressure, low-density lipoprotein, glycated hemoglobin (A1C), diabetes duration, hypertension, albumin excretion rate, glomerular filtration rate, smoking and retinopathy (38,40,41). In addition to the clinical endpoints of CAD, stroke and PVD, subclinical carotid disease may be commonly associated with type 1 diabetes. Compared to age-/sex-matched healthy controls, greater carotid intima-media thickness (IMT) has been observed in studies of children with type 1 diabetes with a mean age as young as 11 years (42–45).
Time course of events
Although CAD rarely presents within the first 20 years of diagnosis, by age 30 years, many individuals will have had type 1 diabetes for 20 years and rates of CVD begin to approach the considered “high-risk” category (46). The recent decline in diabetic kidney disease has not been accompanied by a corresponding fall in CAD rates. Indeed, no temporal decline was noted for the cumulative incidence of MI/CAD death at 20, 25, or 30 years' duration of diabetes in the Pittsburgh EDC, despite at least a 50% decrement of the cumulative incidence of overt nephropathy (31). In fact, nephropathy or microalbuminuria no longer precedes CAD in the majority of cases. In the EDC study, there was no difference in the cumulative incidence of CAD stratified according to year of diagnosis (1950-1980), despite substantial declines in renal failure as well as decline in overall mortality over the same time period (31). The DCCT intensive therapy intervention had a significant impact on the age and the duration of diabetes exposure at onset of CVD, despite the fact that no overt CVD was apparent at baseline (47). Thus, despite the well-recognized increase in CVD risk associated with proteinuria, it clearly explains only a portion of the CVD risk. In the DCCT study, the treatment group effect of intensive treatment therapy on CVD risk persisted after adjustment for microalbuminuria (hazard ratio [HR] 0.62) and albuminuria (HR 0.58), suggesting that, although diabetic kidney disease is important, differences in mean A1C are clearly significant drivers (47). In the same way, only 15% of the Oslo Study population had microalbuminuria, despite the fact that all participants had at least subclinical CAD (48). In the Pittsburgh EDC study, myocardial ischemia by ECG, as the initial manifestation of CAD, was less common and a documented MI as more common in those with prior renal disease compared to those without (49).
Effect of gender
Compared to women without diabetes, women with type 1 diabetes had a 3.5 times higher risk of having coronary artery calcifications (50). While standardized mortality rates from ischemic heart disease were higher in men than women at all ages in the general population, there was no difference in mortality from ischemic heart disease in men and women with type 1 diabetes <40 years of age (36). Men with type 1 diabetes age ≥40 years had a higher mortality rate from CVD than women with type 1 diabetes (51). In a large Norwegian cohort study, mortality rates from ischemic heart disease were higher in women with type 1 diabetes than in men or women without diabetes. However, men with type 1 diabetes had higher mortality rates than women with type 1 diabetes (52). The most recent population-based cohort study showed different results (53). This study found that among those with type 1 diabetes, women had a 2.5 to 3 times higher standardized mortality rate from CVD than men with type 1 diabetes. Although not all the findings are consistent, the common thread in all these studies is that the presence of type 1 diabetes (as well as in type 2 diabetes) seems to dramatically increase the risk for CVD, particularly in women.
Testing for CVD in type 1 diabetes
In the absence of data to the contrary, 1 approach to identifying CVD in patients with type 1 diabetes is to apply the same CAD risk assessment and diagnostic strategies used in type 2 diabetes (see discussion above) or in the population in general (54). This, however, does not support routine CAD screening beyond resting ECGs in patients with diabetes who do not have cardiovascular symptoms or an abnormal ECG, favouring instead global risk factor assessment and treatment.
Patients with type 1 diabetes who have symptoms suggestive of CAD, an abnormal resting ECG or clustering of cardiac risk factors yielding an intermediate or high global risk estimate, acknowledging that risk scores are more or less accurate in type 1 diabetes, should have additional testing for CAD (54,55). For patients able to walk on a treadmill without significant baseline ST segment abnormality (see discussion for type 2 diabetes), exercise treadmill testing remains the first-line diagnostic test due to the high cost efficacy and widespread availability. However, treadmill testing may not be possible due to the burden of peripheral neuropathy, foot pathology, lower extremity amputation and ECG abnormalities as left ventricular hypertrophy in the patient population with type 1 diabetes. Pharmacological stress imaging studies, such as nuclear myocardial perfusion imaging or pharmacological stress echocardiography may be required. Sophisticated testing has been reported in patients with type 1 diabetes. Coronary artery calcium, assessed by CT imaging, is common (56,57) and more frequent in patients with type 1 diabetes than in those without, and is seen at higher rates than in those without diabetes. Progression of coronary artery calcium is reduced by intensive glycemic control (57). The presence of coronary artery calcium is independently associated with increased prevalence of CAD, even after adjustment for traditional risk factors (56) , and test performance in patients with type 1 diabetes is comparable to that of the general population. In the Pittsburgh EDC longitudinal study, 302 adults with type 1 diabetes, with a mean age of 38 years, underwent coronary artery calcium screening. The prevalence of coronary artery calcium was 11% in patients <30 years of age and as high as 88% among those 50 to 55 years. Coronary artery calcium was independently associated with prevalent CAD across the entire cohort, with a stronger graded association in men than in women. While coronary artery calcium assessment has proven to predict subsequent cardiovascular risk in the general population and in cohorts of patients with type 2 diabetes (58) , no data are yet available to determine the utility of coronary artery calcium assessment for risk prediction in type 1 diabetes. Women with type 1 diabetes had just as much coronary artery calcification as men; women without diabetes have less coronary calcium than men (50).