OUP user menu

Clinical studies on coronary revascularization in patients with type 2 diabetes

Koon-Hou Mak, David P. Faxon
DOI: http://dx.doi.org/10.1016/S0195-668X(03)00111-8 1087-1103 First published online: 2 June 2003


With the increasing prevalence, diabetes is rapidly growing into a global public health problem. Cardiovascular disease is a major consequence of this chronic condition, and a critical issue facing physicians worldwide today is choice of coronary revascularization procedures in treating these patients. Since the Bypass Angioplasty Revascularization Investigation (BARI) alert in 1995, there have been several reports on subgroup analysis of clinical trials and registries concerning revascularization among patients with diabetes. Although randomized control studies comparing percutaneous and surgical revascularization procedures in this high-risk group of patients are lacking, this article provides the background for the excess risk and reviews the findings of these investigations. Current strategies to improve outcomes in patients with diabetes undergoing coronary revascularization procedures are discussed.

  • Angioplasty
  • Bypass
  • GP IIb/IIIa inhibitor
  • Stent

1 Introduction

Diabetes is a chronic disease with high cardiovascular morbidity and mortality.1 Approximately 150 million worldwide are suffering from this condition and the number is expected to rise to 300 million by 2025. In 1997, 22 million people in Europe suffer from diabetes and the number is projected to increase to 32.9 million in 2010.2 While only 17 million Americans suffer from diabetes in 1997,3 the total cost burden for this condition was estimated to be $98 billion, with more than half of the expenditure attributed to indirect cost such as temporary or permanent disability, and premature death.4 Of the $44 billion used for direct medical costs, almost a fifth was employed for management of cardiovascular complications. With a million new patients diagnosed with diabetes each year in the United States, the strain on financial and medical resources will continue to increase. Undoubtedly, diabetes has evolved into a major public health issue and is of great concern to both healthcare providers and administrators.

Although there has been considerable improvement in managing patients with coronary artery disease, unfavourable events remained heightened among patients with diabetes.5 For example, mortality was greater among diabetics with acute myocardial infarction, even when they were treated with contemporary fibrinolytic regimen6 or direct angioplasty.7 With the prevalence of diabetes rising, the number of patients with diabetes undergoing revascularization procedures is expected to increase. Unfortunately, the occurrence of adverse outcomes following percutaneous coronary intervention8,9 and coronary artery bypass grafting10,11 is still higher among patients with diabetes. In 1996, the Bypass Angioplasty Revascularization Investigation (BARI)12 revealed the wide gap in cardiovascular outcomes among patients with diabetes and multi-vessel disease randomly assigned to balloon angioplasty or coronary artery bypass grafting. Currently, ∼20% of patients referred for revascularization procedures are diabetics. With a large heterogenous body of information emerging, the triaging process to medical therapy, percutaneous coronary intervention or coronary artery bypass grafting has become increasing complex. Thisarticle reviews the findings of clinical studies on revascularization procedures comparing between patients with and without type 2 diabetes, and various therapeutic options in improving outcomes among this high-risk group of patients undergoing percutaneous coronary intervention.

2 Biological and metabolic abnormalities

Several biological and metabolic abnormalities exist in patients with diabetes rendering them particularly vulnerable to vascular complications. Indeed, even among non-diabetic with acute myocardial infarction, hyperglycaemia portends poor prognosis.13,14 Beyond elevation of blood sugar level, the aberrations of diabetes involve a complex interplay of a variety of processes, including a prothrombotic state, endothelial dysfunction, exuberance of growth factors and matrix formation, which has not been fully understood. Nonetheless, these factors provide a basis to explain the poor outcomes of patients with diabetesundergoing coronary revascularization.

2.1 Haemostatic abnormalities

Clearly, patients with diabetes have an enhanced thrombotic state with impaired fibrinolysis. Platelets are larger with greater number of glycoprotein (GP) Ib and IIb/IIIa receptors,15 and tend to aggregate more easily in an ex-vivo model.16 These aberrations may be related to activation of protein kinase C, reduced nitric oxide, increased active oxygen radicals and abnormal calcium metabolism.17,18 Furthermore, plasma concentrations of fibrinogen,19 coagulation factor VII20 and von Willebrand factor21 are elevated in diabetics, and thrombin activity is increased.22

The imbalance of the fibrinolytic system may also contribute to the problem of restenosis in patients with diabetes. Using samples derived from directional atherectomy, diabetics were found to have considerably higher levels of plasminogen activator inhibitor type I and lower levels of urokinase-type plasminogen activator.23 Taken together, these factors are likely to produce a prothrombotic state in patients with diabetes, and may account for a less favourable outcome.

2.2 Endothelial dysfunction

The important role of a normal functioning endothelium is increasingly recognized in vascular biology. Besides being a barrier between blood and tissue, it is highly metabolically active and regulates blood flow, delivery of numerous substances and cells across this single-layer.24 Notably, impairment of nitric oxide-mediated endothelial relaxation has been described in diabetics prior todevelopment of overt atherosclerosis,25 with production of prostacyclin reduced26 and endothelin-1 increased.27 High sugar level inhibits nitric oxide synthetase activation and increases oxidative stress by generating free radicals.28 Furthermore, acute hyperglycaemia29 and advanced glycation end-products30 have been also shown to impair endothelial relaxation. These factors result in lowered production of nitric oxide and diminished endothelial regenerative capacity.31

2.3 Exaggerated cellular and matrix proliferation

The roles of insulin-mediated mitogens, such as insulin-like growth factor-I (IGF-I), and hyperglycaemia in promoting cellular and matrix proliferation are complicated. Nonetheless, they are pivotal in the enhancing the process of restenosis in diabetics. Together with platelet-derived growth factors, insulin-like growth factor-I is known to stimulate smooth muscle migration andproliferation.32 Hyperglycaemia increases growth factors such as basic fibroblastic growth factor and transforming growth factor-α,33 and synthesis of matrix elements such as collagen type IV, fibronectin, and laminin.34 Correspondingly, lipid abnormalities in patients with diabetes, such as elevated very low-density lipoprotein, enhance the expression of endothelial nuclear factor-κB, adhesion molecules intercellular adhesion molecule-1, vascular cell adhesion molecule-1, tumour necrosis factor-α35 promoting the development of atherosclerosis. Hyperglycaemia also transforms low-density lipoprotein particles to smaller and more atherogenic oxidized or glycated particles which stimulate smooth muscle cell migration and induce apoptosis.36 Furthermore, advanced glycation end products may lead to chronic cell activation, and probably proliferation.37 As neointimal formation following balloon angioplasty is greater among patients with diabetes,38 therapeutic strategies in preventing smooth muscle migration and proliferation are likely to reduce restenosis rates in this high-risk group of patients.

2.4 Insulin resistance

The clinical significance of insulin resistance is gaining prominence, and has been associated with several medical conditions such as polycystic ovarian syndrome and syndrome X. While insulin resistance may precede the onset of type 2 diabetes, its presence portends a heightened risk of occurrence of myocardial infarction39 and stroke.40 Indeed, the United Kingdom Prospective Diabetes Study (UKPDS) showed a reduction in macrovascular events by improving insulin resistance with metformin in obese patients with type 2 diabetes.41 Although the precise mechanisms for the deleterious effects of insulin resistance remain uncertain, excessive free fatty acids are likely to be a major contributor. They reduce production of the nitric oxide and other vasodilatory substances mediated through reactive oxygen radicals.42

Recently, this condition has been shown to be associated with increased neointimal hyperplasia, and restenosis following balloon angioplasty is higher among patients with elevated insulin levels after a glucose load43 Likewise, among patients with hyperinsulinaemia undergoing coronary stenting, there was increased neointimal proliferation in a serial intravascular ultrasound study.44 Interestingly, the exaggerated response to injury was observed only in insulin-resistant and not non-insulin resistant rats.45 The rate of 6-month target revascularization in a large clinical trial on coronary stenting was considerably higher amongpatients with clinical characteristics of insulin resistance (diabetes, hypertension and obesity) compared with those without (16.7% vs 7.5%; p<0.001). This difference persisted after adjusting for baseline characteristics.46

2.5 Burden of coronary atherosclerosis

Although an early angiographic study suggested that coronary atherosclerosis among patients with diabetes was not more diffuse than non-diabetics,47 subsequent investigation showed that they are associated with more extensive and diffuse coronary atherosclerosis.48 In addition, the prevalence of multivessel disease was higher among diabetics with myocardial infarction (65.4% vs 52.2%; p<0.001).6 Conversely, others reported that the extent49 and severity50 of symptomatic coronary atherosclerosis were comparable between patients with and without diabetes. But the incidence of left main disease remained higher in diabetics (13 vs 6%; p<0.01).49 Recently, a more robust autopsy study51 reported that those with diabetes were more likely to have any high-grade (75% vs 55%; p<0.0001) and multivessel high-grade coronary atherosclerosis (58% vs 41%; p<0.001). Lesion morphological characteristics also differed between unstable angina patients with and without diabetes.52 Plaque ulceration and intracoronary thrombus were observed more frequently among diabetics. Taken together, these features mayexplain partly the higher occurrence of ischaemic events and poorer outcomes of patients withdiabetes.

3 Coronary angioplasty

3.1 Balloon angioplasty

Although not widely recognized initially, the unfavourable influence of diabetes among patients undergoing conventional balloon angioplasty has been well-documented. Despite being older, more likely to be female, having a greater extent of coronary atherosclerosis and co-morbid conditions, procedural success rates were comparable between patients with and without diabetes.8,9 However, in-hospital adverse outcomes (Table 1), long-term mortality and need for repeat revascularization procedures were higher among diabetics. When the data from two early studies were put together,8,9 the odds ratio (OR) for a patient with diabetes suffering from death or myocardial infarction following balloon angioplasty was 1.7 times greater than that of a non-diabetic during the hospitalization period. Long-term survival was also considerably lower among patients with diabetes (Fig. 1, Panel A). The probabilities of 9-year survival for patients with and without diabetes were 68.0% (95% confidence interval [CI], 62.9% to 73.2%) and 83.5% (95% CI, 82.0% to 85.0%), respectively (p<0.0001). Furthermore, the proportion of patients who suffered from subsequent myocardial infarction was lower among those without diabetes (p<0.0001) (Fig. 1, Panel B). Similarly, long-term survival was poorer among patients with diabetes undergoing balloon angioplasty in a Dutch centre.53 After 15 years, the cumulative survival rates were 52% (95% CI, 42% to 62%) and 68% (95% CI, 64% to 72%) for patients with and without diabetes, respectively (p<0.0001). Indeed, the presence of treateddiabetes was a predictor for long-term mortality (relative risk, 1.82, 95% CI, 1.16 to 2.86).

View this table:
Table 1

Balloon angioplasty: in-hospital event rates and odds ratio*

Event rates
Event(n=1414)(n=11 133)Odds ratio95% CIp-value
Death14 (1.0)33 (0.3)3.341.79 to 6.23<0.001
MI27 (1.9)154 (1.4)1.380.92 to 2.070.12
Death/MI40 (2.8)185 (1.7)1.701.22 to 2.390.002
CABG35 (2.5)259 (2.3)1.060.75 to 1.510.73
  • * Data derived from Stein et al.8 and Kip et al.9Expressed in numbers with percentages in parenthesis.CI=confidence interval; MI=myocardial infarction; CABG=coronary artery bypass grafting.

Fig. 1

Estimates showing the probability of survival (Panel A, left) and freedom from myocardial infarction (Panel B, right) between patients with (broken line) and without (solid line) diabetes undergoing balloon angioplasty. Data are derived from life tables obtained from the Emory8 and National Heart, Lung, and Blood Percutaneous Transluminal Coronary Angioplatsy9 Registries. Mth=month; FF=freedom from.

Notably, differences in outcomes have beenobserved within the first 6 months after the index procedure. Indeed, angiographic restenosis rate following balloon angioplasty has been reported to be 62% (95% CI, 58% to 66%) in 476 lesions from 377 patients with diabetes,54 which is considerably higher than contemporary balloon angioplasty restenosis rates of 30%. Importantly, total occlusion occurred in 12.6% of the patients, of whom 36% remained asymptomatic. Those with total occlusion were associated with a 6.2% reduction in left ventricular ejection fraction, and may partly account for the poorer outcomes for diabetics following balloon angioplasty. Other independent predictor for restenosis in this study54 included treatment of saphenous vein grafts (OR, 16.6; 95% CI, 2.1 to 131.7), abnormal pre-procedural antegrade flow (OR, 3.2; 95% CI, 1.7 to 6.0), bifurcation (OR, 1.8; 95% CI, 1.0 to 3.2), organ damage (OR, 1.7; 95% CI, 1.1 to 2.7), and post-procedural residual stenosis (OR, 1.03; 1.01 to 1.05).

With greater restenosis rates, the need for coronary artery bypass grafting (p<0.0001) and repeat coronary angioplasty (p<0.0001) procedures were not unexpectedly higher among diabetics (Fig. 2). Between patients with and without diabetes, the estimates for subsequent coronary artery bypass grafting began to diverge after the first 2 years, while the difference for the need for repeatpercutaneous coronary intervention was observed much earlier.

Fig. 2

Estimates showing the probability of freedom from coronary artery bypass surgery (Panel A, left) and repeat coronary angioplasty (Panel B, right) between patients with (broken line) and without (solid line) diabetes undergoing balloon angioplasty. Data are derived from life tables obtained from the Emory8 and National Heart, Lung, and Blood Percutaneous Transluminal Coronary Angioplatsy9 Registries. Mth=month; FF=freedom from.

Several hypotheses have been put forward to explain the poor outcomes of patients with diabetes mellitus undergoing balloon angioplasty. These patients have smaller caliber vessels and probably from negative remodeling, particularly among those treated with insulin.55 In addition, another reason may be the high occlusive restenosis rates54 which is associated with higher mortality among 513 diabetic patients,56 After a meanperiod of 6.5 years, compared with patientswithout restenosis, the adjusted hazards ratio (HR) for total mortality among those with nonocclusive restenosis was 1.6 (95% CI, 1.1 to 2.6; p=0.03), and those with occlusive restenosis was 3.1 (95% CI, 1.9 to 5.1; p=0.0001). Although restenosis was not significantly higher in diabetics, new narrowings were more frequently encountered, particularly in the vessel which has been treated percutaneously.57 Pair cineangiograms of the 248 patients undergoing balloon angioplasty and repeat coronary angiography at a mean period of 250 days later were examined independently for the development of new disease in the coronary arterial tree. New lesions appeared more commonly (OR, 1.7; 95% CI, 1.1 to 2.6) among patients with diabetes (14.8% vs 8.5%; p=0.01). New stenoses appeared most frequently in treated diabetic vessels (16.9%) followed by non-treated diabetic vessels (13.2%), treated non-diabetic vessels (12.7%) and non-treated non-diabetic vessels (7.3%; p=0.009). The OR of new narrowings among treated diabetic vessels was 2.5 (95% CI, 1.3 to 4.7) higher than non-treated non-diabetic vessels. These findings suggested that disease progression was considerably faster among patients with diabetes, particularly when the vessel has been instrumented by angioplasty devices, and likely to also explain the greater occurrence of adverse events following coronary angioplasty for diabetics.

Although the rapid disease progression in diabetes was not shown in the Emory Angioplasty Surgery Trial (EAST),58 the BARI Investigators observed more lesion progression (23.2% vs 17.4%) and development of new lesions (3.5% vs 2.0%) at 5 years among diabetics undergoing balloon angioplasty. Another factor believed to contribute to the poorer outcomes for patients with diabetes undergoing balloon angioplasty was small caliber vessels.8 However, a recent study59 reported that the diabetic status rather than vessel diameter has been reported to be an independent determinant for adverse long-term cardiac outcomes. In addition, event-free survival for diabetic patients with glycosylated haemoglobin <6% was comparable to non-diabetics upto the first 5 years after the index procedure.

3.2 Atherectomy devices

The experience with directional atherectomy has been similar to balloon angioplasty. Of 512 patients treated with directional atherectomy in the Coronary Angioplasty Versus Excisional Atherectomy Trial (CAVEAT-I), there were 95 (18.6%) with diabetes.60 Acute results were comparable between patients with and without diabetes. The 6-month angiographic restenosis rates (59.7% vs 47.4%; p=0.068) and need for subsequent revascularization procedures at 6 months (31.6% vs 24.5%; p=0.19) were marginally higher among those with diabetes. Likewise, the angiographic restenosis rate was higher in diabetics when high-speed rotational atherectomy was used to treat diffuse (∼25mm long) coronary artery disease (72.2% vs 45.5%; p<0.05).61 Importantly, the length of restenosis was considerably longer among patients with diabetes (18.1 vs 11.5mm; p<0.05).

4 Balloon angioplasty and bypass surgery in multivessel disease

Since its introduction, coronary angioplasty has progressed significantly, from treating single proximal lesions to complex multi-lesion, multi-vessel disease, challenging coronary artery bypass grafting as the principal mode of revascularization therapy. In addition to relieving symptoms, coronary artery bypass grafting has been shown to improve survival in certain subsets of patients.62 Although both technologies have been evolving rapidly, investigators were keen to evaluate the outcomes of different revascularization strategies. As such, several registries and clinical trials were established in the late 1980s and early 1990s to compare the efficacy of percutaneous coronary intervention with coronary artery bypass grafting among patients with multi-vessel disease. While most reports demonstrated similar survivalbetween percutaneous coronary intervention and coronary artery bypass grafting, patients with diabetes undergoing balloon angioplasty were consistently showed to suffer from higher mortality. However, there is no prospective randomized trial designed to directly compare these two techniques of coronary revascularization in patients with diabetes.

4.1 Mortality comparison among randomized trials

Among five randomized trials of patients with multivessel disease comparing between coronary angioplasty and coronary artery bypass grafting, there were 627 patients with diabetes.63–66Between a follow-up period of 1 and 8 years, mortality was consistently lower among diabetics undergoing coronary artery bypass grafting (Fig. 3), although the association was less evident when follow-up duration was short. Taken together, mortality was 33.7% (95% CI, 28.3 to 39.1%) among those undergoing coronary angioplasty and 18.9% (95% CI, 14.5 to 23.3%) among those undergoing coronary artery bypass grafting.

Fig. 3

Bar chart showing mortality rates, obtained from randomized trials, of patients with diabetes undergoing coronary artery bypass grafting (CABG) and coronary angioplasty. The number at the bottom of each pair of bars indicates length of follow-up in years (yr). EAST=Emory Angioplasty Versus Surgery Trial; BARI=Bypass Angioplasty Revascularization Investigation; CABRI=Coronary Angioplasty versus Bypass Revascularization Investigation; ERACI=Argentine Randomized Study: Coronary Angioplasty versus Coronary Bypass Surgery in Multivessel Disease.

The BARI Investigators offered some insightinto the higher mortality among diabetic patients undergoing balloon angioplasty.67 Notably, the survival benefit afforded by coronary artery bypass grafting was provided by the use of the internal mammary artery graft, resulting in a striking reduction of cardiac mortality.68 This conduit was believed to be less susceptible to disease progression, and was more likely to be available for myocardial perfusion during episodes of coronary occlusion, even from a remote area, as evident by the protection offered to patients undergoing coronary artery bypass grafting who subsequentlysustained a Q-wave myocardial infarction.69Furthermore, distal dispersion of vasoactive substances may retard progression of native vessel disease. However, the type of revascularization procedure did not influence the occurrence of myocardial infarction, and the rate was comparable to earlier studies.70,71 But diabetic patients with Q-wave myocardial infarction after coronary artery bypass grafting were less likely to die, with an adjusted relative risk of only 0.09 (95% CI, 0.03 to 0.29; p<0.001) compared with those who underwent balloon angioplasty.69 Although the likelihood of death was also lowered for those treated with surgery but did not suffer from a Q-wave myocardial infarction, the estimate was smaller (adjusted RR, 0.65; 95% CI, 0.45 to 0.94; p=0.02). These differences were observed soon after onset of myocardial infarction, and this relationship was stronger among those receiving the left internal mammary artery conduits.

Another reason for the poorer outcome among patients with diabetes treated with balloon angioplasty compared with coronary artery bypass grafting may be related to progression of disease. In BARI, 369 of 914 coronary artery bypass grafting patients and 528 of 915 balloon angioplasty patients who were successfully revascularized and underwent ≥1 follow-up angiography were evaluated for the amount of myocardium at risk after the initial procedure.72 Diabetes was present in 58 and 97 patients treated with coronary artery bypass grafting and balloon angioplasty, respectively. Among patients undergoing coronary artery bypass grafting, there was no increase in the percentage of jeopardized myocardium between those with and without diabetes. Conversely, among patients undergoing balloon angioplasty, those with diabetes had a substantial increase in the proportion of jeopardized myocardium at 1-year (protocol-directed) (42% vs 24%; p=0.05) and 30-month (clinically driven) angiography (63% vs 26%; p=0.01), but not at 5-year (protocol directed) angiography (34% vs 26%; p=0.33).

4.2 Mortality comparison among registries

In contrast to the findings from subgroup analysis from randomized trials, the difference in mortality among diabetic patients undergoing percutaneous coronary intervention or coronary artery bypass grafting was less evident. From a broad period of study extending >20 years, survival data of 8818 patients derived from seven registries (Fig. 4), with follow-up periods ranging from 5 to 12 years, were reviewed.68,73–78 Of these, only two showed increased mortality among diabetic patients undergoing balloon angioplasty.76,77 The overall probability of long-term mortality of 27.8% for the percutaneous strategy compared with 26.3% for surgery. The reason for the discrepancy between subgroup analysis of clinical trials and observational data is unclear. However, coronary artery bypass grafting patients in registries were much sicker than those undergoing percutaneous coronary intervention. In the BARI study,79 1829 patients consented and randomized, and 2010 others with multivessel disease were followed-up in a registry. Among those in the Registry, educational level and quality of life was higher, and they were more physically active and were less likely to smoke cigarettes than randomized patients. Of those receiving revascularization in the BARI registry, almost two-thirds were treated with percutaneous coronary intervention. There was little difference in 7-year mortality between these two therapeutic modalities, including diabetics. Likely patient selection remains as a critical determining factor, and practicing clinicians are able to choose the appropriate revascularization strategy forindividuals with diabetes.

Fig. 4

Bar chart showing mortality rates, obtained from non-randomized studies, of patients with diabetes undergoing coronary artery bypass grafting (CABG) and coronary angioplasty. Duration of follow-up was expressed in years. EMORY=Emory University Hospitals Database;73 LDCMC=Lady Davis Carmel Medical Center Registry;74 DUKE=Duke University Medical Center Database;75 MAHI=Mid America Heart Institute;76BARI-R=Bypass Angioplasty Revascularization InvestigationRegistry;68 NNE=Northern New England Cardiovascular Disease Study Group;77 APPROACH=Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease.78

Despite coronary artery bypass grafting provided a more favourable outcome than balloon angioplasty for patients with diabetes, the risk of diabetics undergoing surgery remained higher than those without diabetes.11,80 However, others81 found comparable short-term survival. Recently, a report of 146 786 patients undergoing coronary artery bypass grafting, of whom 28.4% suffered from diabetes,82 showed that 30-day mortality was higher in those with diabetes (3.7% vs 2.7%), with an adjusted OR of 1.23 (95% CI, 1.15 to 1.32). Not surprisingly, stroke, renal failure and infections occurred more frequently among patients with diabetes. By reducing these peri-operative morbidities, outcomes of patients with diabetes undergoing coronary artery bypass grafting can be further improved.

4.3 Need for repeat revascularization procedures

In the BARI trial,64 revascularization procedures were performed more frequently among patients randomized to percutaneous coronary intervention regardless of diabetic status (Fig. 5). Notably, there was little difference in repeat procedures among patients with and without diabetes in the coronary artery bypass grafting group. In contrast, the rate of repeat revascularization procedureswas significantly greater among diabetics undergoing percutaneous coronary intervention than non-diabetics, with the excess accounted by higher coronary artery bypass grafting rates. Among patients undergoing percutaneous coronary intervention, the odds for a patient with diabetes requiring subsequent coronary artery bypass grafting in next 7 years was 1.9 (95% CI, 1.3 to 2.6) higher.

Fig. 5

Need for subsequent revascularization at 7 years after the index procedures for patients with and without diabetes in the Bypass Angioplasty Revascularization Investigation.64

5 Coronary stenting

5.1 Balloon angioplasty vs coronary stenting in patients with diabetes

While the occurrence of restenosis has been shown to be lower among selected patients undergoing coronary stenting compared with balloon angioplasty,83,84 its benefit among diabetics was less certain. From a total of 555 patients in four reports85–88 (Fig. 6), the angiographic restenosis rate was considerably higher among those treated with balloon angioplasty compared with coronary stenting (59% vs. 29%; p<0.001). The rate of occlusion at 6-month follow-up was also higher for patients treated with balloon angioplasty (4% vs 13%; p<0.005).88 Patients with occlusion were associated with a substantial reduction in left ventricular ejection fraction (9.9±11.2%; p=0.001) which was not observed among those without restenosis or with non-occlusive restenosis. Due to the higher occurrence of occlusive restenosis among those treated with balloon angioplasty, there was a significant fall in left ventricular ejection fraction (2.4±10.9%; p=0.02), while there was little change in this measurement in the stented group. Among those treated with stents, the rates of death or nonfatal myocardial infarction (14.8% vs 26.0%; p=0.02) and repeat revascularization procedures (35.4% vs 52.1%; p=0.001) were lower at 4 years. These benefits translated to a higher event-free survival rates for patients with diabetes treated with stents (58.8% vs 36.9%; p<0.0001). Furthermore, among patients with chronic total occlusion, there was a trend towards higher patency rates for those with diabetes treated with coronary stents compared with balloon angioplasty (96.3% vs 81.6%; p=0.076) at follow-up.89 Although the results are promising, but whether coronary stenting provides similar benefit to diabetic and non-diabetic patients remains uncertain.

Fig. 6

6-month angiographic restenosis rates among patients with diabetes undergoing balloon angioplasty and coronary stenting. Data were derived from van Belle et al.,85 STRESS=Stent Restenosis Studies I and II,86 ISAR-SMART=Intracoronary Stenting or Angioplasty for Restenosis Reduction in Small Arteries,87 van Belle et al.88

5.2 Coronary stenting in patients with and without diabetes

An early observational study,90 using the Palmaz-Schatz stent, reported that restenosis occurred substantially higher among patients with diabetes (55% vs 20%; p=0.001). Despite similar acute results, late loss was significantly greater among patients with diabetes (1.66 vs 1.23mm; p=0.04), and the authors attributed this finding to increased intimal hyperplasia, which was subsequently validated by intravascular ultrasound examination.38

While there are no randomized trials, several observational and case-control studies provided some insight on the impact of coronary stenting among patients with and without diabetes. From five studies,95,90–93 with >4800 patients, angiographic restenosis rate was consistently higher among diabetic patients undergoing coronary stenting (Fig. 7, Panel A). Despite using different stents over a broad range of patients, the overall 5 to 6-month angiographic restenosis rates were 36.8% and 26.3% for patients with and without diabetes, respectively (p<0.001), translating to an increased OR of 1.6 (95% CI, 1.4 to 1.9). In fact, the presence of diabetes was an independent predictor for the occurrence of death, myocardial infarction or stroke at 9 years in patients with single-vessel disease.94

Fig. 7

Angiographic restenosis (Panel A, left) and clinical restenosis (Panel B, right) among patients with and without diabetes treated by coronary stenting. Data derived from Carozza et al.,90 Lau et al.,91 van Belle et al.,85 Elezi et al.,92 and Schofer et al.,93 for Panel A; and Elezi et al.,92 Abizaid et al.,97 Marso et al.,98 and Carrozza et al.,99 for Panel B.

Similar to balloon angioplasty, diabetic patients had a greater propensity (OR, 2.1; 95% CI, 1.5 to 3.0) to suffer from total occlusion (6.4% vs 3.1%) after coronary stenting.85,92,93 Furthermore, the interaction between heightened mortality andocclusive restenosis following balloon angioplasty may be operative, accounting for the pooreroutcomes.56

Unlike balloon angioplasty, vessel recoil is almost eliminated by coronary stenting, the underlying mechanism of renarrowing is largely attributed exaggerated neointimal hyperplasia. Therefore, the reference diameter may be an important predictor and smaller vessel were more likely to develop restenosis in multivariate analysis.92 After coronary stenting, restenosis rates were higher than non-diabetic patients only in vessels <3mm in diameter (44% vs 23%; p=0.002) and not vessels ≥3mm (18% vs 15%).95

Other investigators96 have observed a dissociation between luminal diameter stenosis and clinical outcomes. Angiographic restenosis, generallydefined as luminal diameter stenosis ≥50%, may have little clinical impact, and rates of adverse events may be more relevant. Based on a pool of ∼11,000 patients obtained from 4 reports92,97–99 (Fig. 7, Panel B), the composite unfavourable event rate, consisting of the occurrence of any of the following at 6 months or 1 year: death, myocardial infarction or repeat revascularization, was substantially higher among patients with diabetes (21.1% vs 15.3%; p<0.001), with an OR of 1.5 (95% CI, 1.3 to 1.7). Notwithstanding the differences in patient groups and use of intravascular ultrasound for stent deployment, the rates of individual component of the composite endpoint remained higher for diabetics. Importantly, at 1 year, patients with diabetes were 2.0 (95% CI, 1.5 to 2.8) and 1.7 (95%CI, 1.3 to 2.2) times more likely to die (6.4% vs 3.2%; p<0.001) or suffer from myocardial infarction (7.2% vs 4.5%; p<0.001), respectively.92,97 Although repeat revascularization procedures were performed more frequently among diabetics, notably there was little difference between the rates of non-target lesion revascularization (8.3% vs 7.2%; p=0.28).92,97 Conversely, in another report on balloon angioplasty,57 the rate of disease progression in treated and non-treated vessels were substantially higher among diabetics.

Better glycaemic control reduces blood thrombogenicity100 and its significance was explored in 151 diabetic patients undergoing coronary stenting.101 Glycosylated haemoglobin level was significantly higher among those with restenosis (7.5% vs 6.8%; p=0.012). Indeed, OR for the occurrence of restenosis was 0.4 (95% CI, 0.2 to 0.8) when glycaemic control was good (<6.5%) compared to those who were poor (>7.5%). Other factors, including cholesterol levels, type of diabetic treatment, lipoprotein (a) and body mass index were not predictors for restenosis. However, glycaemic control hadlittle impact on long-term adverse clinical outcomes in another study.102 But the definition of control less stringent (good, <8.0%; poor, >10%), and may have accounted for the lack of significant association.

6 Coronary stenting and bypass surgery in multivessel disease

Both coronary stents and coronary artery bypass grafting, with the use of arterial conduits, have improved outcomes in patients with coronary artery disease. Although there are randomized studies comparing between these 2 revascularization strategies, their findings have been conflicting.103–105 The Arterial Revascularization Therapies Study (ARTS) demonstrated comparable mortality between patients undergoing coronary stenting (2.5%) and bypass surgery (2.8%) at 1 year.103 Conversely, the second Argentine Randomized Study: Coronary Angioplasty versus Coronary Bypass Surgery in Multivessel Disease (ERACI II)104 reported a higher likelihood of death among those undergoing coronary artery bypass grafting (7.6% and 5.4%; p=0.017) after a mean follow-up period of 18.5 months. On the other hand, the Stent or Surgery Study (SoS) found that mortality was significantly higher among those treated with percutaneous coronary intervention (4.5%) than coronary artery bypass grafting (1.6%) (HR, 2.91, 95% CI, 1.29 to 6.53; p=0.01) after a median follow-up period of 2 years.105 The wide discrepancies in relative and absolute mortality among these three studies may be a reflection on patient selection and procedural expertise. Not surprisingly, patients receiving stents were more likely to require repeat revascularization procedures (16.6% vs 7.3%; p<0.001).103,104

In the ERACI II study,104 diabetes was not a significant predictor for 30-day adverse cardiovascular outcomes in a multivariable model. However, this may be attributed to the small number of patients with diabetes. The influence of diabetes on long-term outcome was not presented. In contrast, the diabetic cohort was described in the ARTS trial.106 Among those treated with stents, mortality was considerably higher among patients with diabetes (6.3% vs 1.6%; p=0.013) (Fig. 8). Regardless of diabetic status, the rate of repeat revascularization was higher in the stented than coronary artery bypass grafting patients (16.6% vs 7.3%; p<0.001).103,104 While these procedures were performed more frequently among diabetic patients receiving stents by 1 year, the difference was not statistically significant. However, using a multivariable model, the patients with diabetes was more than twice as likely to experience any adverse cardiovascular events at 1 year.

Fig. 8

Bar chart showing 1-year clinical outcomes in the Arterial Revascularization Therapy Study (ARTS) Trial. Only significant p-values are shown. CABG=coronary artery bypass grafting. ▪, Death; □, Myocardial infarction; Embedded Image, Stroke; Embedded Image, Coronary artery bypass grafting; Embedded Image, Coronary angioplasty

Adjunctive medical therapies differed between the ERACI II and ARTS trial. Notably, the proportions of stented patients treated with abciximab in these two trials were 28% and 3%, respectively. This important difference may account, at least partly, for the variation in outcome. Others107,108 have attributed the excess in adverse events amongdiabetic patients with multivessel disease undergoing percutaneous coronary intervention, at least in part, to incomplete revascularization. Whether this explanation applies to the comparison between coronary stenting and coronary artery bypass grafting remains controversial. A similar analysis for the SoS study is planned.

Novel techniques in preventing myocardialinjuries during coronary revascularization procedures may further improve outcomes. The use of beating heart surgery109 may enhance the safety and efficacy of coronary artery bypass grafting. Reports focussing on cellular protective agents have been encouraging. The sodium-hydrogenexchanger110 is activated during ischaemia leading to intracellular accumulation of hydrogen ions, and subsequent injury and necrosis. Cariporide (HOE642A), a potent inhibitor of the exchanger, was extensively studied in the GUARDIAN (GUARd During Ischemia Against Necrosis) trial.111 There were 3439 and 2918 patients undergoing high-risk percutaneous coronary intervention and coronary artery bypass grafting, respectively. Diabetes was present in 33% and 47% of the percutaneous coronary intervention and coronary artery bypass grafting groups, respectively. Patients were randomly assigned to 3 dosing regimens and infusion of cariporide commenced prior to the procedure, and treatment discontinued after 2 to 7 days. There was a trend towards lower adverse cardiovascular events in the low-dose percutaneous coronary intervention group, but a 25% relative reduction in death or myocardial infarction at 36 days in the coronary artery bypass grafting group. There was no information on patients with diabetes.

7 Glycoprotein IIb/IIIa blockade

7.1 Balloon angioplasty

The complementary interaction between abciximab and coronary angioplasty was explored in the Evaluation in PTCA to Improve Long-term Outcome with Abciximab GP IIb/IIIa Blockade (EPILOG) study.112 Of 2792 patients, 638 (23%) had diabetes. Compared with placebo, bolus and infusion of abciximab reduced the occurrence of death or myocardial infarction at 6 months (HR, 0.36, 95% CI, 0.21 to 0.61). Surprisingly, the amount of benefit afforded by abciximab was marginally higher than patients without diabetes (HR, 0.60, 95% CI, 0.44 to 0.82). Interestingly, the investigators observed a relationship between heparin dose and outcome. Among patients with diabetes, the rate of death, myocardial infarction or urgent repeat re-intervention was slightly lower among thosereceiving standard-dose (100U/kg) than low-dose heparin (70U/kg) (23.1% vs 31.4%; p=0.074). This finding may reflect upon the need for greater anticoagulation for patients with diabetes undergoing balloon angioplasty, even in the setting of potent antiplatelet inhibition.

7.2 Coronary stenting

With the favourable results from coronary stenting and GP IIb/IIIa blockade, the synergistic combination will potentially provide far better outcomes. In the Evaluation of Platelet IIb/IIIa Inhibitor for Stenting Trial (EPISTENT), there were 491 (20.5%) patients with diabetes, a pre-specified subgroup.113 The 6-month occurrence of death, myocardial infarction or target vessel revascularization was lowest among those treated with stent plus abciximab (13.0%) compared with balloon angioplasty plus abciximab (23.4%) or stent only (25.2%) (p=0.005). Compared with diabetic patients treated with stent alone, those receiving stent plus abciximab have lower incidence of death or myocardial infarction (6.2% vs 12.7%; p=0.041). Indeed, by pooling the results of three large clinical trials114 comparing abciximab with placebo among patients with diabetes undergoing percutaneous coronary intervention using a variety of devices, 1-year mortality was substantially lower among those receiving GP IIb/IIIa blockade (2.5% vs 4.5%; p=0.031).

In the EPISTENT study113 , target vessel revascularization was reduced among diabetic patients treated with stent plus abciximab (8.1% vs 16.6%; p=0.021), to a rate comparable to non-diabetics treated with stent plus abciximab (8.8%). Of 149 patients (30.3%) who underwent 6-month angiography, net gain was considerably greater among those treated with stent plus abciximab compared with stent alone (0.88 vs 0.55mm; p=0.011), with a marginally lower loss index (0.40 vs 0.60; p=0.061). Although abciximab has not been shown to prevent angiographic restenosis,115 its effect on patients with diabetes is less certain. Currently, there is an on-going randomized placebo-control trial,Abciximab in Stenting Inhibits restenosis Among Diabetics (ASIAD), evaluating the role of coronary stent plus abciximab in preventing angiographic restenosis in 260 patients with diabetes. The effect of abciximab on late adverse events in 570 patients with diabetes undergoing non-acute myocardial infarction percutaneous coronary intervention was also reviewed based on the Mayo Clinic Registry.116 While the rates of 30-day mortality (0.6% vs 3.0%; p=0.03) and repeat percutaneous coronary intervention (0% vs 1.1%; p=0.03) were lower among those treated with abciximab, there was littledifference in individual or composite unfavourable cardiovascular events at 1-year, even after adjusting for baseline characteristics.

The Do Tirofiban and ReoPro Give SimilarEfficacy Trial (TARGET)117 compared the efficacy of tirofiban and abciximab to prevent ischaemic complications among patients undergoing coronary stenting. Overall, at 30 days, the composite event of death, myocardial infarction or urgent target vessel revascularization occurred more frequently among those randomized to tirofiban (7.6% vs 6.0%; p=0.038). Of 4809 patients enrolled, there were 1117 (23.2%) who had diabetes. In this subgroup of patients, there was little difference in the incidence of the composite endpoint (tirofiban, 15.7%, abciximab, 16.9%; p=0.61), with a comparable rate of target vessel revascularization at 6 months (tirofiban, 9.5%, abciximab, 11.1%; p=0.366).118 Mortality at 1-year was also similar (tirofiban, 2.1%, abciximab, 2.9%; p=0.436). In contrast, administration of eptifibatide was not shown to be superior to placebo among diabetic patients undergoing coronary stenting.119 The lack of benefit for eptifibatide may be related to an unexpectedly low risk group of patients with diabetes in the study.

8 Current strategy and future directions

The modality of coronary revascularization for patients with diabetes and multivessel disease remains controversial. Several issues were not addressed by contemporary clinical studies, such as the type of treatment, adequacy of glycaemic control, aggressive management of other atherosclerotic risk factors, adjunctive therapy with statins and angiotensin converting enzyme inhibitors, and status of distal vessel bed. With rapidly evolving technologies in the arena of interventional cardiology and cardiac surgery, most techniques employed by the investigators were outmoded by the time the results became widely available. Notwithstanding these limitations, coronary artery bypass grafting was shown to provide better outcomes. However, the synergism of coronary stenting and GP IIb/IIIa blockade may afford similar results between patients with and without diabetes. Currently, lesion morphological characteristics are likely to play an important role in determining the type of revascularization procedure. Patients with discrete lesions in large caliber vessels may be suitable for percutaneous coronary intervention, particularly when theproximal left anterior descending artery is not significantly diseased. Furthermore, percutaneous coronary intervention may be the procedure of choice when the left internal mammary artery is not available as a conduit or when the patient has several other co-morbid medical conditions. However, the use of other arterial conduits may provide comparable outcomes.120,121 Conversely, patients with diffuse narrowings in small caliber vessels and poor left ventricular function should undergo coronary artery bypass grafting. Among patients whereby concomitant cardiac procedures are required, such as valve surgery, coronary artery bypass grafting is more appropriate.

With the number of patients with diabetes increasing worldwide, there are several on-going studies evaluating numerous strategies in their optimizing outcomes. The advent of drug-eluting coronary stents has changed the entire landscape of interventional cardiology with the promise of no restenosis122 The encouraging results of the sirolimus-eluting stent was amplified in the much larger RAVEL (Randomized Double-Blind Study with the Sirolimus-Eluting Bx Velocity (Cordis, Warren, NJ) Balloon-Expandable Stent in the Treatment of Patients with De Novo Native Coronary Lesions) Study, consisting of 238 patients.123 At 6 months, the angiographic restenosis rates for the stents with and without drugs were 0% and 26%, respectively (p<0.0001), with a 1-year event-free survival of 94.2% and 71.2% (p<0.0001), respectively.124 Of 44 patients with diabetes, 19 were treated with the sirolimus-eluting stent and 25 with uncoated stents. The restenosis rates were 0% and 42% (p<0.0001), respectively. The late loss for patients with diabetes was comparable to non-diabetics (0.08 vs −0.01mm, respectively). In the larger SIRIUS trial (Sirolimus-Coated Bx Velocity Balloon-Expandable Stent in the Treatment of Patients with De Novo Coronary Artery Lesions), the overall results on 1053 patients showed an in-segment restenosis rates of 8.9% and 36.3% (p<0.0001) for stents with and without sirolimus.125 There were 279 patients (26.5%) with diabetes, and thein-segment restenosis rate was also lower among those treated with sirolimus-eluting stent (17.6% vs 50.5%; p<0.001). Despite using sirolimus-eluting stents, patients with diabetes continued to have higher restenosis rate than those without. It remains unclear if this finding was dose-related. However, the relative reduction of restenosis was comparable to non-diabetic patients. The FREEDOM (Future Revascularization Evaluation in Patients with Diabetes Mellitus: Optimal Management of Multivessel Disease) study, a prospective multicenter international randomized trial comparing the drug-eluting stent with bypass surgery among 1500 patients with diabetes, is being planned. An European registry, ARTS-2, will compare the outcomes of patients with multivessel disease treated with sirolimus-eluting stents and coronary artery bypass grafting.

With the favourable results from the use ofmetformin in preventing myocardial infarction in the UKPDS,41 other investigators focussed on improving insulin sensitivity as a key interventional strategy for patients with type 2 diabetes undergoing coronary revascularization. A novel receptor, peroxisome proliferator-activator receptor-γ (PPAR-γ), which regulates gene expression, appears to be a promising approach to improve outcomes of patients with type 2 diabetes undergoing coronary revascularization.126 Agonists such as thiazolidinediones bind to peroxisome proliferator-activator receptor-γ and promote insulin-mediated glucose transport into adipose and skeletal muscles, lowering hyperglycaemia and improving insulin resistance. Recently, this group of pharmacological agents has been shown to reduce smooth muscle cell proliferation,127 and rosiglitazone (Avandia®, GlaxoSmithKline, Philadelphia, PA) was shown to reduce neointimal hyperplasia in patients with type 2 diabetes undergoing coronary stenting.128 Currently, the PPAR (peroxisome proliferator-activator receptor-γ agonist for the prevention of lateadverse events following percutaneous revascularization) is being conducted to evaluate the efficacy of rosiglitazone, compared with placebo, in preventing 1-year adverse cardiac events in 200 obese patients with type 2 diabetes or hypertension. As advanced glycation end products alter the compliance of vessel walls, and responses toinjury,37 delinking sugars and amino groups is an attractive approach to restore normal vascular characteristics. A thiazolium derivative, ALT-711 (3-phenyl acyl-4,5-dimethylthiazolium bromide) breaks down these cross-linkages129 and reduce wall stiffness.130 The use of this novel agent may provide a rational alternative to improve outcomes of patients with diabetes undergoing coronary revascularization. Finally, the BARI-2D trial aims to enroll 2600 patients with type II diabetes evaluating the role of revascularization and glycaemic control in a 2X2 matrix design. Patients are to be randomized to medical therapy or revascularization, employing either percutaneous coronary intervention or coronary artery bypass grafting at the investigator's discretion. As for type of glycaemic control, patients are randomly allocated to insulin-sensistizing agents, such as thiazolidinedione or biguanide, or insulin-providing drugs, such as sulphonylurea. Physicians worldwide are eagerly awaiting the desperately needed results of these new studies to provide more data and better manage this extremely high-risk group of patients.


View Abstract