Worldwide, it is estimated that the total number of people with diabetes will increase from 171 million in 2000 to 366 million in 2030.1 In the US, a total of 23.6 million people have diabetes, representing 7.8% of the population, and the prevalence of diabetes is on the rise, due to an increasingly sedentary lifestyle, increasing obesity and an ageing population. These statistics do not account for the approximately 57 million people with impaired fasting glucose.2
Previous data have clearly demonstrated diabetes to be a highly significant risk factor for coronary artery disease (CAD). Diabetic patients without a prior diagnosis of CAD are as likely to have an incident myocardial infarction (MI) as non-diabetic patients with prior MI.3 CAD is the leading cause of death among diabetic patients,4 with atherosclerosis accounting for 65–80% of all deaths among this population.5 Although there has been a 49% reduction of cardiovascular events over the last 50 years in diabetic patients, due in part to more aggressive preventative and revascularisation strategies, the incidence of adverse cardiovascular events is two-fold greater in diabetics compared with non-diabetics.6
Because diabetic patients are at a higher risk of CAD and adverse cardiovascular events, they constitute an ideal population in whom screening for CAD could aid in risk stratification and prevention of cardiovascular events. Symptomatic diabetic patients are clearly at high risk and require screening (in almost all cases, we favour anatomical delineation of CAD in these patients). However, there are conflicting opinions in terms of the value of screening asymptomatic patients with diabetes for CAD. The most recent consensus statement from the American Diabetes Association (ADA) does not recommend routine screening of asymptomatic diabetic patients for CAD, unless medical treatment goals cannot be met or there is a strong clinical suspicion of very high risk.7 Patients with co-existing risk factors, such as advanced age, significant peripheral vascular disease or advanced diabetes (such as patients with other end-organ manifestations of diabetes), are typically sufficiently high risk to merit screening. However, recognising which asymptomatic patients without these conditions are at high risk is often challenging. Patients may have a limited functional capacity, which may lead a physician to question whether they are truly asymptomatic. Additionally, the prevalence of silent ischaemia is greater in diabetic patients, and patients may present with a major cardiac event without prior warning.8 Diabetic patients with silent ischaemia have higher cardiac mortality than patients with symptoms – 26 versus 9% – most likely because they present later and revascularisation is delayed.9
A recent randomised trial, Detection of Ischemia in Asymptomatic Diabetics (DIAD), highlighted some of the concerns of screening in asymptomatic diabetic patients. In this trial, 1,123 asymptomatic diabetic patients were randomised to screening with stress singlephoton- emission computed tomography (SPECT) or no screening and followed up for a mean of 4.8 years.10,11 Overall, patients in DIAD were at very low risk of cardiac events, with a cumulative event rate of 2.9% over the study period (0.6% per year), a mean glycated haemoglobin (HbA1c) of 7% and a 9% prevalence of peripheral arterial disease. These characteristics suggest that higher-risk patients were not enrolled in the trial; for example, in the group who received screening SPECT scans, 22% had silent ischaemia, with only 6% of patients having moderate or large perfusion defects. There were no differences observed between the two randomised groups in terms of all-cause death, cardiac death, MI and unstable angina. Of note, however, is that even among this low-risk group of patients, when outcomes were assessed in the screened group according to the amount of quantifiable ischaemia on stress testing, patients with moderate or large defects had a six-fold greater risk of events compared with patients with normal scans or small defects seen on SPECT. These findings are consistent with those of a pilot study of asymptomatic high-risk diabetics, in whom there was a 22% risk reduction in cardiac events among patients screened with exercise electrocardiogram (ECG) and stress echocardiography.12
As a result of these considerations, asymptomatic diabetic patients are often screened non-invasively for CAD, and patients without demonstrable ischaemia are thought to be at low to intermediate risk of subsequent cardiovascular events. On the other hand, the management of asymptomatic diabetic patients with demonstrable ischaemia on stress testing can often be vexing to the clinician, who must weigh the up-front assumed risk of invasive procedures (both diagnostic angiography and/or revascularisation) against the putative benefits of a reduction in ischaemic burden. The presence of myocardial ischaemia predicts poorer long-term survival in diabetics than non-diabetics, and the higher mortality associated with diabetes may be mediated by ischaemia to a large extent, with limited differences observed in subsequent mortality among diabetic and non-diabetic patients without ischaemia.13 High-risk findings on SPECT imaging have been observed in 18% of asymptomatic diabetics with a subsequent annual mortality rate of 5.9%, as opposed to a mortality rate of 2.9% in patients with low-risk SPECT findings.14 Of note, among patients with high-risk SPECT findings in this study who subsequently underwent coronary angiography, 61% had high-risk CAD. These data are consistent with other studies demonstrating a strong association between the amount of ischaemia on stress SPECT and subsequent cardiovascular events in patients with stable CAD.15,16
On demonstration of significant myocardial ischaemia, the majority of patients with diabetes (even those who are asymptomatic) should typically undergo further risk stratification through diagnostic coronary angiography. Some would argue whether this is necessary for patients with minimal or small amounts of ischaemia, and in these patients there is undoubtedly an emerging role for non-invasive anatomical risk assessment, for example through coronary computerised tomography angiography (CTA). Clearly, the decision to proceed with diagnostic angiography is a clinical decision that should take into account not only the amount of ischaemia but also the baseline clinical risk (including co-morbidities), the functional capacity of the patient and other high-risk features of the stress test. However, it should be noted, that current clinical paradigms, as well as clinical trials randomising patients to revascularisation versus medical therapy, have typically required anatomical delineation of the extent of CAD in order to adequately assess patient risk in individuals with demonstrable ischaemia. Individuals undergoing diagnostic angiography with low-risk CAD are often managed medically, whereas patients with more critical CAD are often presented with options for revascularisation.
The clinical justification for revascularisation of patients with significant CAD is largely based on prior data associating revascularisation of at-risk myocardial territories with improved clinical outcomes. In an analysis of patients included in the Cedars Sinai database, revascularisation has been shown to be associated with a reduction in cardiac mortality among patients with a high-risk stress SPECT scan, defined as greater than 10–20% ischaemic myocardium – a finding even more pronounced among diabetic patients.15 In this database, mortality rates were observed to increase progressively, with greater amounts of ischaemia among patients undergoing medical therapy but not in patients referred for revascularisation, reinforcing the ‘protective’ benefit of revascularisation for patients with the highest ischaemic risk. Further data from the nuclear sub-study of the Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation (COURAGE) trial have corroborated a greater degree of ischaemia reduction with revascularisation compared with medical therapy, with a clear association between ischaemia reduction and improved clinical outcomes.16 Additionally, medically managed patients with residual ischaemia on SPECT scans were more likely to subsequently receive revascularisation. In another series, revascularisation provided a 43% relative risk reduction in mortality in asymptomatic diabetics with a high-risk SPECT scan.17 Finally, in the Asymptomatic Cardiac Ischemia Pilot (ACIP) and Swiss Interventional Study on Silent Ischemia Type II (SWISSI II) studies, ischaemia-guided therapy with revascularisation in addition to medical therapy was shown to reduce cardiac death and MI compared with medical therapy alone.18,19
Despite these prior observational data associating revascularisation of appropriately selected patients with improved clinical outcomes, recent randomised studies (COURAGE and Bypass Angioplasty Revascularization Investigation 2 Diabetes [BARI 2D]) have called into question the benefit of revascularisation in such patients, suggesting that a strategy of medical therapy should be the first line of therapy for stable CAD.20,21 However, a critical factor that is often missing in the discussion of these trials relates to the characteristics of the patients enrolled in these studies. For example, while the COURAGE trial did not demonstrate an improvement in clinical outcomes with percutaneous coronary intervention (PCI) compared with medical therapy, despite the low-risk features of patients enrolled in the trial, the nuclear sub-study of the trial clearly demonstrated an association between ischaemia reduction and improved clinical outcomes, a finding that was more pronounced among patients undergoing revascularisation with PCI.16 More recently, the results of the BARI 2D trial have been published, and are additionally instructive. BARI 2D randomised diabetic patients with angiographically demonstrated CAD to a strategy of prompt revascularisation (PCI or coronary artery bypass grafting [CABG], based on the discretion of the treating physicians) versus medical therapy. In the trial, the prompt revascularisation strategy was not superior to medical therapy in preventing so-called ‘hard’ cardiovascular events. Interestingly, among patients deemed suitable for randomisation to PCI or medical therapy, there was no significant difference between revascularisation and medical therapy, while among patients with more extensive CAD deemed suitable for randomisation to CABG versus medical therapy, the rate of subsequent cardiovascular events was significantly lower with revascularisation compared with medical therapy.
BARI 2D was a test of prompt revascularisation versus initial intensive medical therapy in patients who did not require a “definite need for invasive intervention as determined by the attending cardiologist.” Thus, in a sense, BARI 2D was a study of the utility of prophylactic revascularisation in patients with diabetes deemed suitable for either management approach. The finding that prompt revascularisation appeared to benefit a subset of enrolled patients – namely, those with more complex disease who were referred for CABG – is evidence in support of the merit of revascularisation for prevention of hard clinical outcomes even in appropriately selected patients with stable forms of CAD. It is noteworthy that a significant improvement in composite death/MI/stroke in CABG-treated patients in BARI 2D was observed in a population that included a <20% prevalence of proximal left anterior descending artery (LAD) stenosis, a high prevalence of only double-vessel disease, minimal or absent angina in most patients and a mean ejection fraction (EF) >55%, despite a dilution of treatment effect resulting from a very high 42% rate of cross-over to revascularisation within the intensive medical therapy arm of the trial. BARI 2D therefore demonstrates the vital role of prophylactic revascularisation for selected patients with diabetes, mild or no symptoms and stable CAD without traditional ‘high-risk’ anatomy.
Among diabetic patients undergoing revascularisation, the choice between PCI and CABG is a complex and individualised decision based on myriad factors, including the clinical scenario, baseline patient risk of each procedure, the patient’s ability to comply with pharmacological treatment regimens, extent of coronary disease, lesion-specific risks of restenosis, suitability of distal anatomical targets for bypass grafting and, importantly, patient preference. While a complete discussion of the choice of PCI versus CABG for patients with diabetes is beyond the scope of this article, several points are worth consideration.
First, although many have interpreted BARI 2D as a comparison between PCI and CABG revascularisation in patients with diabetes, it is critical to note that this trial was not a randomised comparison of these two strategies. Randomisation between medical therapy and prompt revascularisation was stratified by the up-front decision to treat with PCI or CABG, but the actual decision in terms of the mode of revascularisation was not randomised. As a result, patients referred for CABG rather than PCI within this study were more likely to have disease of the proximal LAD, a greater number of lesions requiring intervention and a greater myocardial jeopardy score, all likely indicators of more myocardial ischaemia. Unfortunately the prevalence, severity and prognostic impact of ischaemia in BARI 2D were not reported. In light of these findings, we believe that the most important message from BARI 2D is not to differentiate among forms of revascularisation, but rather that revascularisation can reduce future cardiovascular events among diabetic patients with more complex forms of coronary disease.
Furthermore, even for diabetic patients with multivessel disease, two recent randomised trials (Coronary Artery Revascularization in Diabetes [CARDIA] and SYNergy between percutaneous coronary intervention with TAXus and cardiac surgery [SYNTAX]) have demonstrated similar one-year rates of death and MI and reduced strokes with PCI utilising drugeluting stents compared with CABG, particularly in those patients with less complex disease.22,23 Further data from these trials are emerging in terms of the most appropriate means of selecting patients for one strategy compared with another, as these trials have confirmed data from earlier trials demonstrating increased rates of repeat revascularisation in diabetic patients undergoing PCI compared with CABG, but no significant differences in hard clinical end-points.24 Five-year data from a patient-level analysis of four trials comparing multivessel PCI utilising bare-metal stents versus CABG (importantly, excluding older studies performed in the era of balloon angioplasty without stent implantation) have demonstrated similar rates of death and MI with both approaches to revascularisation.25
We expect more definitive evidence to emerge from the ongoing National Institutes of Health (NIH)-sponsored Future REvascularization Evaluation in Patients with Diabetes Mellitus: Optimal Management of Multivessel Disease (FREEDOM) trial, a multicentre, open-label, prospective, randomised superiority trial of PCI versus CABG in diabetic patients in whom revascularisation is indicated. Patients will be randomised on a 1:1 basis to either CABG or multivessel stenting using drug-eluting stents and observed at 30 days, one year and annually for up to five years.26 The primary end-point of the trial is a composite of death, non-fatal MI and stroke.
In summary, the management of asymptomatic diabetic patients presents a challenge to the clinician. Clinical risk assessment is paramount when deciding whom and when to screen for CAD, and when to proceed with anatomical delineation of CAD (diagnostic coronary angiography). Of note, it is important to recognise the prevalence of silent ischaemia in this population since it predisposes to higher rates of cardiac death and MI. Multiple studies have demonstrated the association between greater ischaemic burden and adverse outcomes, and patients at the highest ischaemic risk stand to gain the most from revascularisation. Any revascularisation strategy has to be considered against the background of optimal medical therapy, and choosing between CABG and PCI is a decision that must be individualised to each patient.