Current Cardiology Reports

, Volume 12, Issue 2, pp 140–146

Screening Asymptomatic Patients with Type 2 Diabetes Mellitus for Coronary Artery Disease: Does It Improve Patient Outcome?


    • Department of CardiologyGeisinger Medical Center
  • Vasken Dilsizian
    • Department of Diagnostic Radiology and Nuclear MedicineThe University of Maryland Hospital and School of Medicine

DOI: 10.1007/s11886-010-0091-z

Cite this article as:
Shirani, J. & Dilsizian, V. Curr Cardiol Rep (2010) 12: 140. doi:10.1007/s11886-010-0091-z


The increasing global burden, the reported high prevalence of rapidly progressive coronary artery disease (CAD), and the atypical nature of CAD presentation in type 2 diabetes mellitus have encouraged development of strategies for detecting occult CAD in this population. Several recent prospective studies have addressed the value of screening for CAD in asymptomatic diabetic patients. The overall message of these studies is that despite detection of silent ischemia in a notable proportion of these patients, the dynamic nature of myocardial ischemia, the prohibitive cost of screening all asymptomatic patients, and the proven efficacy of primary preventive strategies would mandate implementation of better clinical risk stratification strategies for identifying at-risk individuals. Questions still remain as to what best strategy would allow proper patient selection through logical stepwise approaches to screening and whether that would alter patients’ outcome when added to rigorously implemented primary preventive measures.


Coronary artery diseaseDiabetes mellitusDetection of ischemia in asymptomatic diabetes (DIAD)Myocardial perfusion imagingScreeningPrognosis

Clinical Trial Acronyms


Action to Control Cardiovascular Risk in Diabetes


Action in Diabetes and Vascular Disease: Preterax and Diamicron-MR Controlled Evaluation


Bypass Angioplasty Revascularization Investigation in Type 2 Diabetics Trial


Diabetes Control and Complications Trial


Detection of Ischemia in Asymptomatic Diabetics


Multicenter Trial of Elective Revascularization in Patients With Diabetes Mellitus and Mild Anginal Complaints


Milan Study on Atherosclerosis and Diabetes


National Health and Nutrition Examination Survey


United Kingdom Prospective Diabetes Study


Veterans Affairs Diabetes Trial


Type 2 diabetes mellitus (DM) is a global health care issue with increasing incidence, prevalence, and cost [1]. It is a well-recognized coronary artery disease (CAD) risk factor, and most patients with DM will have CAD and suffer from its complications [2]. The outcome of patients with DM and no known CAD is similar to nondiabetic patients with prior myocardial infarction, and the overall prognosis of myocardial infarction is worse in those with than without DM [3]. This poor clinical outcome has encouraged a search for practical approaches to early detection of CAD in asymptomatic adults with DM. This article summarizes the recent developments in the assessment of CAD risk, extent, and outcome in asymptomatic diabetic individuals.

The Global Burden of Diabetes

The prevalence of DM, arguably the fastest growing public health issue in the adult population, was estimated to be 4.0% in 1995 and was projected to rise to 5.4% by the year 2025 [1]. This meant that in 1995 there were 135 million diagnosed adult diabetic individuals in the world, and that this number would rise to 300 million by 2025 [1]. The projected number for the year 2030 is 366 million [4]. Diabetes is especially prevalent in the developed countries, such that nearly 10% (>20 million) of the adults in the United States suffer from the disease [5•]. Data from the NHANES 1999 to 2002 showed that the prevalence of DM increases sharply with age and that the disease remains undiagnosed in a sizable number of individuals. From 1990 to 2001, the prevalence of diagnosed DM in the United States increased 61% [6]. Based on the 1984 and 2004 data, the total prevalence of diagnosed DM in the United States will reach 12% in 2050, with the largest increase among the older age groups [7]. The National Diabetes Statistics Fact Sheet indicates that in 2005, 1.5 million new cases of DM were diagnosed in the United States among those ≥20 years of age. It should be noted that in 2006, 57 million Americans (26% of the population) were believed to have qualified for the diagnosis of prediabetes [7].

Diabetes and CAD

At least 65% of diabetic adults die of cardiovascular causes and death rates from heart disease are two to four times higher in those with than without DM based on the data from the US National Institute of Diabetes and Digestive and Kidney Diseases. Diabetes significantly increases the risk for developing cardiovascular disease. The latter is believed to be the major reason for a nearly 8-year shorter life expectancy of the diabetic patients [8]. Among the cardiovascular conditions, CAD is the most prevalent in the diabetic patients and for the large part accounts for the increased mortality in these patients. DM is believed to have mitigated the recently observed trends toward reduction in age-adjusted CAD mortality in the United States [9].

Much has been learned about the role of hyperglycemia in the pathogenesis of diabetic complications from the DCCT. It was shown that intensive blood sugar control in type 1 DM can have long-lasting beneficial effects through “metabolic memory” [10•], and that 10 years after the initiation of such therapy cardiovascular complications can be reduced by 57% [11]. Unfortunately, such information is presently less convincing for type 2 DM despite the fact that it represents more than 90% of all cases of diabetes.

Three major studies of tight glycemic control in type 2 DM produced conflicting results on its impact on macrovascular complications [1214••], and one [12] suggested that very tight control (hemoglobin A1c < 6.0%) may in fact be detrimental to those with pre-existing cardiovascular disease and long duration of DM. However, 10-year follow-up results from the UKPDS suggests that early, intensive glycemic control does reduce the long-term cardiovascular complications in type 2 DM [15]. Despite an ongoing controversy [16], it was also shown that longstanding (>5 years since diagnosis) type 2 DM can be regarded as a myocardial infarction equivalent because the long-term cardiovascular mortality is similar in diabetic patients without prior myocardial infarction and nondiabetic patients with pre-existing myocardial infarction [17]. The latter supports the previous positions taken by the American and European experts [18, 19].

Understanding the relationship between CAD and DM is further complicated by the fact that the former often fails to produce recognizable symptoms in the diabetic patient. This is despite the fact that atherosclerosis tends to be more severe in diabetic compared with nondiabetic patients, as shown in a case-control autopsy study of more than 2000 individuals [20]. In this study, nearly 75% of diabetic patients without clinical CAD had high-grade coronary stenoses and more than 50% had multivessel CAD [20]. Comparative clinical diagnostic studies provided clear evidence for the high prevalence of coronary atherosclerosis in asymptomatic individuals with DM [2126]. For example, coronary artery calcification has been frequently detected by CT in diabetic adults and is found to be severe in 20% of such individuals [27]. Stress radionuclide myocardial perfusion imaging has demonstrated that the extent of coronary calcification directly correlates with the presence and severity of perfusion defects in that setting. Thus, 32% of asymptomatic diabetic patients with moderate coronary calcification were shown to have myocardial perfusion defects, and the corresponding number for those with severe coronary calcification has been 60% [27].

Prospective Studies of Screening for CAD in DM

Three recent prospective studies examined the value of CAD screening in asymptomatic individuals with type 2 DM (Table 1) [2833]. In the MiSAD study [28], each of the seven participating hospitals in the Milan area recruited 160 adults (age, 40–65 years) with type 2 DM and no microvascular or macrovascular diabetic complications, including history of CAD or other cardiac conditions. On entry to the study, each subject underwent maximal exercise electrocardiographic (ECG) study followed by exercise thallium scintigraphy only in those with positive or equivocal result. Of the 1,120 enrolled patients, 195 were excluded because of detection of exclusion criterion after enrollment, abnormal cardiac tests indicating prior undetected CAD, or failure to complete the protocol. The remaining 925 subjects completed the study protocol.
Table 1

Clinical, diagnostic, and follow-up data from prospective studies examining CAD screening in asymptomatic individuals with type 2 diabetes mellitus


MiSAD [28, 29]

North London [30]

DIAD [31, 32]









Mean age, y

54 ± 6

53 ± 8

61 ± 7

Men, %




Duration of diabetes, y




Hypertension, %




Smoker, %




Total cholesterol, mg/dL

220 ± 42

192 ± 37

198 ± NA

Hemoglobin A1c,%

7.3 ± 1.5

8.2 ± 1.7

7.1 ± 1.6

Insulin use allowed




Screening test

Maximal ETT, selective MPI

CAC score, selective MPI


Screening mode




Silent ischemia, %




Follow-up duration, y




Adverse events, % per year

 Normal study




 Abnormal study




Adverse events, % per year





 Not screened




CAC coronary artery calcium, CAD coronary artery disease, DIAD Detection of Ischemia in Asymptomatic Diabetics, ETT exercise treadmill test, MiSAD Milan Study on Atherosclerosis and Diabetes, MPI myocardial perfusion imaging, NA not available

Overall, the exercise ECG test was abnormal in 12%, of whom 53% also had perfusion abnormalities on thallium scintigraphy. The study could not provide an overall estimate of myocardial perfusion defects in asymptomatic patients with DM, as only those with abnormal stress ECG underwent perfusion imaging. However, the study was able to emphasize the significance of resting and exercise ECG abnormalities in this population. On the basis of these screening results, 52 subjects with abnormal scintigraphy were evaluated by a cardiologist and 22 underwent coronary angiography [29]. Obstructive CAD (>50% luminal narrowing) was present in all patients and five had three-vessel disease, of whom four underwent coronary artery bypass surgery. Percutaneous coronary intervention was performed on four subjects. No further cardiac workup was offered to those with normal exercise ECG or normal thallium scintigraphy at baseline. Over a 5-year follow-up period, fatal and nonfatal myocardial infarction occurred in 3% of the 638 subjects with normal baseline exercise ECG tests, 2% of the 45 subjects with normal thallium scans, and 2% of the 52 subjects with abnormal thallium scans [29]. These rather low event rates were attributed to early detection and active treatment of CAD in this highly selected, relatively low-risk group that also underwent aggressive risk factor modification [29]. A strategy of stress ECG testing followed by myocardial perfusion imaging was proposed as a potentially cost-effective strategy for CAD screening in asymptomatic diabetic adults.

In the North London study, asymptomatic patients with uncomplicated (no microvascular or macrovascular complications) type 2 DM were recruited from four medical clinics in the United Kingdom [30]. Coronary calcium scores were calculated by electron beam CT (Agatston units) in 510 subjects who fulfilled all admission criteria and were categorized into minimal (<11), mild (11–100), moderate (101–400), severe (401–1,000), and extensive (>1,000) calcification. All patients with coronary calcium scores greater than 100 [n = 127 (93%)] and a random sample of those with scores ≤100 [n = 53 (14%)] underwent 99mTc-sestamibi dipyridamole single photon emission computed tomography (SPECT) myocardial perfusion imaging. Myocardial perfusion defects were identified in 57 of the 180 patients (32%) and were mostly reversible (89%). There was a clear direct relationship between the coronary artery calcification and presence of perfusion defects, which ranged from none in those with minimal to 71% in those with extensive calcification. Likewise, moderate-to-large perfusion defects were more often seen in those with calcium scores greater than 100 than in those with calcium scores ≤100 (18% vs 32%). Abnormalities other than perfusion defects (eg, stress-induced ST-segment depression, transient left ventricular dilatation, and baseline left ventricular dysfunction) were also identified in 12% of the patients. By multivariable logistic regression analysis, coronary calcium score was the sole predictor of myocardial perfusion abnormality (P < 0.001). Attending physicians were made aware of the study’s findings.

At a median follow-up duration of 2.2 years, 20 major adverse cardiovascular events (eg, cardiac death, nonfatal myocardial infarction, acute coronary syndrome, non-hemorrhagic stroke, or late revascularization) occurred. All events occurred in patients with abnormal perfusion scans (n = 16) or high coronary calcium score despite normal perfusion scans. No cardiovascular events occurred in subjects with minimal coronary calcification. Fifteen of the 20 events occurred in patients with severe or extensive coronary calcification. The area under the receiver operator curve was 0.92 for coronary calcium score for predicting cardiovascular events, which performed better than the traditional clinical risk score models. By multivariable regression, coronary calcium score and the extent of myocardial ischemia on perfusion scan were the only independent predictors of outcome. This study concluded that coronary calcification is present in nearly one half of asymptomatic individuals with uncomplicated type 2 DM as predicted by established CAD risk factors, and that a synergistic relationship existed between coronary calcium score and myocardial perfusion defect in predicting future adverse cardiovascular events. The findings suggested that stepwise use of coronary calcium scoring and myocardial perfusion imaging would provide a potentially effective strategy for screening asymptomatic adult diabetic patients for CAD.

The DIAD study is the only randomized, multicenter (14 centers in the United States and Canada) trial of screening for CAD in asymptomatic adults (age, 50–75 years) with DM [31]. The study was initiated in July 2000, enrolled 1,123 patients by August 2002, ended data collection for primary outcome measure in September 2007, and was completed by January 2008. Inclusion criteria consisted of late-onset (30 years of age or more) DM, no history of ketoacidosis, and no symptoms, signs, history, or ECG evidence of CAD. Overall, 66% (1,123/1,700) of all eligible patients agreed to participate in the study. Of these, 522 completed the 99mTc-sestamibi stress adenosine SPECT imaging arm protocol and 562 were assigned to follow-up alone. SPECT images were interpreted as normal or abnormal (eg, ischemia, scar, or mixed defects) and small, moderate, or large. Increased radiotracer lung uptake, transient stress-induced left ventricular dilation, and reduced left ventricular ejection fraction (<45%) were also considered abnormal. The prevalence of silent myocardial ischemia in subjects randomized to screening was 22%, of which only 6% was quantified as moderate or large [31]. Additionally, 6% of the patients had only ischemic ECG changes during adenosine administration. Of the original 522 patients, 69% (n = 358) had repeat stress imaging 3 years after enrollment [32]. New ischemia was identified in 10% (n = 28) of those with normal scans at baseline, whereas 79% (n = 56) of those with abnormal baseline scans showed resolution of ischemia [32]. Patients who had undergone coronary revascularization were excluded from this analysis. The resolution of ischemia on myocardial perfusion scan occurred irrespective of the magnitude of the initial perfusion defect and directly related to primary preventive efforts, including higher use of statins, aspirin, and angiotensin-converting enzyme inhibitors compared with baseline [32]. At 5 years, the overall cardiac event rate (cardiac death and nonfatal myocardial infarction) was only 2.9% (0.6% per year) and was not significantly different in the two groups (2.7% vs 3.0%, P = 0.73) [33]. Secondary outcome (heart failure, unstable angina, or coronary revascularization) event rates were also low (0.9% per year) and not different between the two groups. However, among screened patients, those with moderate-to-large perfusion defects had significantly higher adverse event rates (2.4% per year) compared with those who had normal or mildly abnormal myocardial perfusion scans (0.4% per year; P = 0.001).

Lessons Learned from CAD Screening in Asymptomatic Diabetic Patients

Despite obvious limitations in the design of the above studies (particularly the inadequate sample size and lack of control over post-test therapeutic decisions), several important lessons have been learned. The first and foremost conclusion is that a distinct group of asymptomatic diabetic patients exist with a low prevalence of silent myocardial ischemia and low cardiovascular event rates (Table 1). These patients may not be comparable to historical groups in retrospective studies that have reported high prevalence of silent myocardial ischemia due to less favorable risk factor profiles or due to less optimal clinical management or both. In the MiSAD study [28], only 7% of diabetic patients who were capable of completing a maximal treadmill exercise had myocardial ischemia even though the perfusion imaging was limited to those with ischemic ECG changes during exercise. Although perfusion defects were present in 32% of those tested in the North London study [30], they were mostly preselected by having at least moderate coronary calcification and 50% could not exercise at low levels on treadmill. In the DIAD trial [31], 22% showed myocardial perfusion defects, of which only 6% were moderate or large. An important additional finding in the latter study was the highly dynamic nature of silent ischemia, as evidenced by the high rate of resolution of myocardial perfusion defects on repeat imaging even in those with moderate-to-large perfusion abnormalities at baseline [32]. The resolution of myocardial ischemia among patients in the DIAD study paralleled the use of preventive medications, indicating that limitations of coronary artery flow reserve (coronary endothelial dysfunction and left ventricular hypertrophy) and not obstructive atherosclerosis may have been responsible for some myocardial perfusion abnormalities [34•]. The latter underscores the need for early institution of effective preventive measures in diabetic patients [35]. Although in the DIAD study myocardial perfusion imaging was able to discriminate those at higher risk for cardiac events based on the size of the perfusion defect (2.4% for those with moderate-to-large vs 0.4% for those with small defects; P = 0.001), nevertheless this distinction was not associated with overall difference in outcome. It should be noted that the overall prevalence of moderate-to-large defects was only 8% in the DIAD study.

Because none of the studies had complete control over the post-test treatment strategies, they cannot provide a final answer as to whether a stepwise (conditional) or collective (unconditional) test-treatment strategy would have had any significant impact on patients’ outcomes. In MiSAD [28, 29], only patients with positive myocardial perfusion scans were placed on aspirin and were referred for cardiology evaluation, and only 22 of the latter 52 subjects underwent coronary angiography. In the North London study [30], the attending physicians were informed of the results of the tests but the final management was left to their discretion. In the DAID study [33], only 5 of 33 (15%) patients with moderate or large defects underwent early coronary angiography, and the rates of overall coronary angiography (14% vs 12%) and revascularization (5.5% vs 7.8%) were not different in screened and not screened subjects. Within these recognized limitations of the aforementioned studies, however, it appears that the use of a lower-cost gatekeeper study (treadmill ECG test or coronary calcium scoring) did not offer additional benefit relevant to clinical outcome of the patients. It is also unlikely that a conditional test-treatment strategy would be more cost effective in this patient population [36].

One important issue regarding the interpretation of the results of the above prospective studies is that the exclusion criteria comprised almost all the features that are recognized as CAD risk factors in patients with DM. From a public health perspective, it then appears that widespread CAD screening of this low-risk asymptomatic group of diabetic patients would be of no significant benefit and will be outweighed by the prohibitive cost of the diagnostic tests. This is further enforced by the results of the BARI-2D trial, which indicated no survival advantage for revascularization as opposed to medical therapy in patients with DM [37••].

Finally, the limitations of the above studies compel us to search diligently for diagnostic algorithms that allow proper identification of diabetic patients at high risk of cardiovascular complications. In addition to symptoms [38•], factors such as age, peripheral vascular disease, carotid artery disease, peripheral neuropathy, autonomic dysfunction, family history of early CAD, longer duration of DM (>10 years), renal insufficiency, adequacy of glycemic control [39•], number of uncontrolled coronary risk factors, and certain genetic traits [40] may predispose diabetic patients to cardiovascular complications. Inclusion of these features into futures studies in which contemporary medical therapies are administered in a controlled manner can provide the necessary information for proper diagnostic evaluation of at-risk individuals.


DM is associated with a heightened risk of CAD and its complications. Recent trends in application of preventive measures may have improved cardiovascular outcomes of the adult diabetic patients [41]. The latter may have influenced the results of the recent prospective studies that have investigated the value of preemptive diagnostic testing for detecting occult CAD in individuals with type 2 DM. Further studies are needed to identify the subset of adult diabetic patients who would most benefit from early identification of CAD by noninvasive testing.


No potential conflicts of interest relevant to this article were reported.

Copyright information

© Springer Science+Business Media, LLC 2010