Featured Article

Diabetes is a major public health problem. The incidence of diabetes among adults in the United States is increasing at epidemic proportions, with almost half of the U.S. states reporting a 6% incidence of diabetes mellitus. The prevalence of diabetes increases with age in both men and women. It peaks by 75 years of age when 21% of men and 17.5% of women have diabetes.¹

Coronary artery disease (CAD) is the most frequent cardiovascular complication of diabetes. It occurs 2–3 times more often in patients with diabetes than in those without diabetes.² Patients with diabetes and no history of CAD have an incidence of cardiovascular death equivalent to non-diabetics who have a history of myocardial infarction (MI). Almost 75% of those with diabetes ultimately die of coronary artery disease (CAD).³ Diabetics have an increased mortality in the setting of both ST segment elevation MI and non-ST elevation acute coronary syndromes. Even diabetic patients who have no history of CAD have an increased risk of cardiovascular death, equivalent to those of non-diabetics who have survived a MI.^4–6

The basis of this excess mortality and morbidity in patients with diabetes secondary to CAD is not completely understood. Most of the classic coronary risk factors are over-represented in diabetics. However, they only explain a small fraction of the excess CAD. Severe hyperglycemia and a long duration of disease enhance the risks of fatal and nonfatal ischemic events, but the effect is quantitatively small as compared to the overall CAD excess.^7,8 Although several autopsy studies reveal severe and diffuse atherosclerotic involvement of the coronary vasculature in diabetics, negative reports are well represented.^9–16

Current understanding of in-vivo diabetic atherosclerosis is based solely on angiography. However, coronary angiography is limited to the display of the luminal contour and is therefore a poor technique to assess atherosclerotic plaque burden, morphology and distribution. Intravascular ultrasound (IVUS) is a more sensitive technique to assess plaque calcification, quantify plaque burden and distribution.^17–23 An IVUS study reported by our group suggests that diabetics with symptomatic CAD have plaque characteristics at the lesion site similar to non-diabetics, but they had a higher plaque burden, resulting in smaller luminal volumes in their diseased coronary arteries when a three dimensional reconstruction of the artery was performed.²⁴

Coronary Revascularization In Diabetics
Despite coronary stenting and newer pharmacological adjunctive therapies, the morbidity and mortality in diabetics after PCI and surgical revascularization remains high.^25–32 No discussion about revascularization in diabetics is complete without referring to the Bypass Angioplasty Revascularization Intervention (BARI) trial.33 In the BARI trial, there was no difference in the survival of patients undergoing percutaneous coronary intervention (PCI) or coronary artery bypass graft (CABG) at 7-year follow up. In the subset of diabetics, the 7-year survival was much lower with percutaneous transluminal coronary angioplasty (PTCA) compared to CABG. However, further analysis revealed that diabetic patients receiving the left internal mammary artery (LIMA) graft were the only patients who benefited significantly from CABG compared to the PCI patients.33 The BARI registry did not show any significant differences (14.4% PTCA versus 14.9% CABG) in 5-year mortality with either approach.34 The Duke database28 as well as the ARTS trial35 did not reveal any difference in 5-year mortality of patients undergoing multi-vessel PCI versus CABG (Table 1).

The reason for the increased morbidity and mortality of diabetics in the BARI trial is not known. However, it is known that diabetics have extensive and severe CAD, which predisposes them to plaque rupture, intra-coronary thrombosis and an increased incidence of coronary events.^36–38 In addition, diabetics have a higher rate of atherosclerotic disease progression. Diabetics have a 22% increase in the number of new narrowings on follow up angiography 1 month after successful angiography compared to 12% in non-diabetic patients.³⁹ Altered platelet function and metabolic milieu contribute to the worse outcome of diabetics.^37,38 Diabetics also have a higher rate of restenosis post PCI. CABG, especially the IMA graft, provides new conduits and a more effective safety net to protect the myocardial beds from acute ischemic syndromes caused by ulcerated plaques.

Strategies to Optimize Percutaneous Revascularization Outcomes
New developments in interventional cardiology have made the results of the BARI less relevant. The strategies to optimize outcomes of diabetics undergoing PCI include those targeting the lesion site and those targeting the diseased diabetic vasculature. The strategies aimed at the lesion site include stenting, Gp IIb/IIIa antagonists, intra-coronary radiation and drug-coated stents. Strategies aimed at the diseased diabetic vasculature include lipid-lowering therapy, role of RAGE, PPAR agonists and control of hyperglycemia.

Targeting the lesion site
A. Stenting. Stents prevent vascular recoil and remodeling. Diabetics, despite stenting, have more recoil and more intimal hyperplasia, which result in a higher rate of restenosis. In addition, diabetics have smaller luminal area in their vessels. Van Belle⁴⁰ in 1997 reported that though diabetics have increased rates of restenosis and reocclusion after PTCA, stenting leads to improved results similar to non-diabetics. Recently, the same authors reported these encouraging results to be maintained 3 years after stenting.⁴¹ The Arterial Revascularization Therapies Study (ARTS) trial³⁵ revealed that diabetics undergoing multi-vessel PCI had similar mortality to patients undergoing CABG. Results from the Argentine Randomized Trial of Percutaneous Transluminal Coronary Angioplasty Versus Coronary Artery Bypass Surgery in Multivessel Disease (ERACI 1)^{42 and ERACI 2 trial,}43 both trials of multi-vessel stenting compared to CABG did not reveal any significant difference of in-hospital, 30-day and 1-year freedom from MI or survival. However, diabetics had a higher requirement of repeat procedures compared to non-diabetics.

B. Gp IIb/IIIa antagonists. Diabetics demonstrate increased platelet adhesion and aggregation. Increased platelet aggregation has been shown to correlate with increased cardiovascular events in diabetics. Diabetics also have larger platelets and contain up to 26% more Gp IIb/IIIa receptors.⁴⁴ Increased glycation of Gp IIb/IIIa receptors has been shown to increase platelet aggregation. Increased platelet dependent thrombin generation and impaired platelet-mediated vasodilatation in diabetics has also been reported.⁴⁴

GPIIb/IIIa antagonists improve outcomes in patients with acute coronary syndromes (ACS) with or without PCI. Diabetic patients undergoing PCI derive an even greater benefit with the use of Gp IIb/IIIa antagonists. A meta-analysis of 6 trials of ACS revealed that the composite of death and MI was reduced by Gp IIb/IIIa antagonists in diabetics at 30 days. There is a significant diabetes treatment interaction on 30-day mortality with the use of Gp IIb/IIIa antagonists.⁴⁵

The abciximab in the Evaluation in PTCA to Improve Long-Term Outcome With Abciximab GP IIb/IIIa Blockade (EPILOG) trial reduced death/MI in diabetics as much as in non-diabetics. The Evaluation of Platelet IIb/IIIa Inhibitor for Stenting (EPISTENT) trial demonstrated that abciximab in conjunction with stenting equalize the outcome of diabetics, similar to non-diabetics with stenting. The EPIC trial, which used abciximab with percutaneous transluminal coronary angioplasty (PTCA) alone, did not demonstrate any beneficial effect in terms of death, MI and target vessel revascularization (TVR) in diabetics. The beneficial effect of abciximab in diabetics was seen only in conjunction with stenting as was seen in the EPISTENT trial.⁴⁵ A substudy of the Platelet Receptor Inhibition in Ischemic Syndrome management in Patients Limited by Unstable Signs and Symptoms (PRISM-PLUS) revealed that the use of tirofiban plus heparin in diabetics reduced the composite outcome of death and MI. Eptifibatide lead to similar results in the Platelet glycoprotein IIb/IIIa in Unstable angina: Receptor Suppression Using Integrilin Therapy (PURSUIT) and Enhanced Suppression of the Platelet IIb/IIIa Receptor with Integrilin Therapy (ESPRIT) trials.⁴⁶ The do Tirofiban And Reopro Give similar Efficacy outcome Trial (TARGET) trial compared the use of abciximab versus tirofiban in patients with ACS. There was no difference in the composite of death/MI/TVR or death or MI or TVR in diabetics with the use of either of the above agents at 6 months or up to 1 year.⁴⁷ A meta-analysis of 3 abciximab trials EPIC, EPILOG and EPISTENT revealed that abciximab reduced the mortality in diabetics (from 4.2% to 2%) but not in non-diabetics at 1 year.⁴⁴ The use of Gp IIb/IIIa antagonists equalizes the outcome of PCI in diabetics similar to non-diabetics.

C. Drug-coated stents. Multiple approaches to restenosis inhibition have been aimed at reducing endothelial injury and proliferation and promoting migration and healing of the endothelial lining after PCI. The use of drug-coated stents, and the agents targeting inflammation and inhibition of proliferation, has become clinically pertinent. The cell cycle specific agents (paclitaxel and rapamycin), and the cell cycle nonspecific agents (actinomycin D) have both shown promise.

None of the individual Cypher or Taxus trials were designed to look specifically at the diabetic patients. However, when you integrate the results of all Cypher trials (n = 292) {SIRIUS (n = 131), E-SIRIUS(33), C-SIRIUS(12), DIRECT(70), SVELTE(27) and RAVEL(19)}, the in-stent binary re-stenosis was reduced by 79% (50.6% in the bare-metal stent arm to 5.7% in the Cypher stent arm p < 0.0001), and the in-segment restenosis was reduced by 80% (52.5% in the bare-metal stent arm to 11.8% in the Cypher stent arm, p < 0.0001). The target lesion revascularization was reduced by 74% (22.3% in the bare-metal stent arm to 5.8% in the Cypher stent arm, p<0.0001) and the target vessel revascularization was reduced by 63% (24% in the bare-metal stent arm to 8.9% in the Cypher stent arm p < 0.0001). In the DIABETES study, in-lesion late loss was assessed at 9 months in 160 diabetic patients with 80 patients each being randomized to bare-metal stent and sirolimus-eluting stent arm. The in-stent restenosis at 9 months was 4.9% versus 31% (p = 0.001), in-lesion restenosis was 7.7% versus 33% (p = 0.001). In the (Formulations of a polymer-based paclitaxel-eluting stent (TAXUS) trials (TAXUS II SR, TAXUS II MR, TAXUS IV SR, TAXUS VI MR) a total of 458 patients had diabetes out of 2289 patients enrolled in the trials. In the Taxus trials, the late loss was 1.03 mm versus 0.36 mm in the bare-metal stent versus the Taxus stent. In non-insulin-dependent diabetic patients, the in-stent binary restenosis was reduced from 32.9% (control) to 4.4% (Taxus), a 87% reduction (p = 0.0001), in-segment binary restenosis was reduced by 82% from 37.5 to 6.6%( p = 0.0001). In 207 patients with diabetes, the True-Life TAXUS registry, the TLR, TVR and non-TVR re-PTCA was 6, 11 and 10%, respectively.

Despite very impressive reductions in restenosis, the advent of drug-eluting stents has still not eliminated diabetes as an independent risk factor for restenosis and adverse events after PCI. The restenosis, both in-stent and in-segment, remains much higher in diabetics compared to non-diabetics. In patients treated with oral hypoglycemic agents and/or insulin compared to non-diabetics, in the Cypher stent trials, the late loss was 0.26 mm compared to 0.15 mm. In patients with diabetes treated with oral agents and receiving a Cypher stent the in-stent re-stenosis was 4.8%, and 9.3% in insulin-requiring diabetics as compared to 2.4% in the non-diabetic patients. In patients with diabetes treated with oral agents and receiving a Cypher stent the in-lesion re-stenosis was 9%, and 22.7% in insulin requiring diabetics as compared to 5.3% in the non-diabetic patients. In patients with diabetes treated with oral agents and receiving a Cypher stent, the 9-month TLR was 4.5%, and 10.1% in insulin-requiring diabetics, compared to 2.7% in the non-diabetic patients. In patients with diabetes treated with oral agents and receiving a Taxus stent, the in-stent late loss was 0.36 mm, and 0.33 mm in insulin-requiring diabetics, compared to 0.36 mm in non-diabetic patients. In patients with diabetes treated with oral agents and receiving a Taxus stent, the in-stent re-stenosis was 4.4%, and 6.7% in insulin-requiring diabetics compared to 4.9% in the non-diabetic patients. In patients with diabetes treated with oral agents and receiving a Taxus stent, the in-segment re-stenosis was 6.6%, and 8.9% in insulin-requiring diabetics, compared to 8.1% in the non-diabetic patients. In patients with diabetes treated with oral agents and receiving a Taxus stent, the 12-month TLR was 6.2%, and 5.6% in insulin-requiring diabetics, compared to 4.6% in the non-diabetic patients. At the present moment, there is no convincing evidence to establish the comparative efficacy of sirolimus, compared to paclitaxel-eluting stents in diabetic patients because of insignificant numbers of diabetic patients in pivotal trials. There are multiple ongoing studies, which will specifically address the question of reduction of restenosis in diabetics with drug-coated stents.⁴⁸

As the incidence of restenosis is reduced, and the native diabetic vasculature is known to be severely diseased, it has been proposed by some authorities that it may be time to consider intimal remodeling therapy⁴⁹ in patients with a high risk of coronary events like diabetics. This would consist of stenting the entire diseased diabetic vessel rather than stenting a severely stenotic segment, which is currently done and called focal plaque sealing.⁵⁰ This would raise interesting possibilities about the cost and safety of such procedures as well as the long-term outcomes of patients undergoing intimal remodeling therapy.

D. Brachytherapy. At present, there is data from many trials showing that brachytherapy is very effective in reducing restenosis. The clinical trials, which used brachytherapy for in-stent restenosis, lead to almost a 60–70% reduction.⁵¹ The SCRIPPS trial⁵² revealed that the use of radiation was even more effective in diabetics as compared to non-diabetics. The Gamma Radiation at One-Year Follow-up for Treating Clogged Stents in Coronary Arteries (GAMMA 1) trial⁵³ also showed that though non-diabetics had a 33% reduction in restenosis with the use of brachytherapy, diabetics had a 52% reduction in restenosis with the use of gamma radiation. Summarizing the data of the diabetics from all the above radiation trials, it can be seen that diabetics had a 75% reduction in TLR, a 68% reduction in TVR, and a 76% reduction in restenosis with the use of intra-coronary brachytherapy, compared to placebo.⁵¹

Strategies Targeting the Vasculature
A. AGE, RAGE and PPAR agonists. Diabetes is an insulin-resistant/deficient state. Insulin resistance, in turn, leads to hypertension, obesity, hyperinsulinemia, hypertriglyceridemia, small dense LOW-DENSITY (LDL), low high-density (HDL) and hypercoagulability, which promote atherosclerosis. Advanced glycation end products induce inflammation, promote cytokine release, and lead to smooth muscle proliferation and increased NF kappa B activity, which leads to increased extracellular matrix and intimal hyperplasia. In addition, advanced glycation end products increase monocyte chemotaxis via the advanced glycation end product (AGE) receptor and thereby promote atherosclerosis. The study of peroxisome proliferator-activated receptor (PPAR) agonists⁵⁴ is a rapidly evolving field. PPAR alpha agonists are well known (omega 3 fatty acids and fibrates). The PPAR gamma agonists are insulin sensitizers such as rosiglitazone and pioglitazone. PPAR alpha and PPAR gamma agonists both influence cardiovascular disease directly or indirectly through their regulation of anti-inflammatory and anti-atherosclerotic gene targets.⁵⁴ Currently, there are multiple ongoing studies trying to determine the role of the PPAR agonists.

B. Lipid-lowering Therapy. When looking at primary and secondary intervention trials, fenofibrate led to a 23% reduction in events in diabetics. Among the secondary intervention trials:
1. Cholesterol And Recurrent Events (CARE) using pravastatin led to a 25% event reduction in diabetics;
2. The Long-term Intervention with Pravastatin in Ischemic Disease study (LIPID) led to a 19% event reduction in diabetics;
3. The Scandinavian Simvastatin Survival Study (4S) using simvastatin led to a 55% event reduction in diabetics;
4. The U.S. VA HDL-C Intervention Trial (VA-HIT) using gemfibrozil led to a 24% event reduction in diabetics.
In the angiographic analysis of the Desmoteplase in Acute Ischemic Stroke (DIAS) study, fenofibrate led to a 40% reduction in minimum lumen diameter, 42% reduction in percent stenosis and 25% reduction in mean segment diameter, compared to a placebo in diabetic patients. Therefore, lipid-lowering therapy is even more effective in diabetics with CAD.^55–58

Diabetes is now considered a CAD equivalent because diabetics need the same intensity of lipid-modifying therapy as non-diabetics with CAD. Diabetics have small dense LDL particles, thus the severity of diabetes associated dyslipidemia is underestimated by standard LDL measurements. The ADA currently recommends that diabetics have a LDL < 100, HDL > 45 and TG < 200. The primary treatment strategy is to reduce LDL, the secondary treatment strategy is to raise HDL and lower TG if > 200. Several experts now suggest that the therapeutic target for LDL be reduced to < 85 mg/dl and HbA1c be reduced to < 6.5 mg/dl to optimally manage diabetic dyslipidemia.⁵⁶

Several ongoing studies are addressing the utility of statins versus fibrates or a combination thereof in the treatment of diabetic dyslipidemia. The three multi-center trials to address this question were the Fenofibrate Intervention and Event Lowering in Diabetics (FIELD) study; the Collaborative Atorvastatin Diabetes (CARDS) study and the Lipids in Diabetes Study (LDS). LDS had to be terminated early due to the withdrawal of cerivastatin from the market. The literature on statin-fibrate combination for diabetic dyslipidemia is limited. The combination of medications has complementary modes of action: the HMGCoA reductase (LDL effect) of statins and the PPAR effect (TG and HDL effect) of fibrates. Studies using simvastatin and bezafibrate combination reduced total cholesterol by 23%, LDL by 29%, TG by 42% and increased HDL by 25% as compared to the atorvastatin–bezafibrate combination which reduced total cholesterol by 37%, LDL by 46%, TG by 50% and increased HDL by 22%. The combined treatment (atorvastatin-fenofibrate) reduced the 10-year probability of MI from 21.6 to 4.2%. Therefore, it seems that the statin-fibrate combination provides additional benefit above either agent.^55–58

C. Glycemic Control and CAD in Diabetics. The United Kingdom Prospective Diabetes Study (UKPDS) suggested that intensive treatment of diabetics led to a 25% reduction in microvascular complications including retinopathy and nephropathy (p = 0.0099), 16% reduction in MI (p = 0.052, ns). Random assignment to either insulin or sulfonylurea drugs was not associated with any increase in cardiovascular disease complications. Metformin monotherapy led to a 39% reduction in myocardial infarction (p = 0.01) in the obese diabetic patients. Recently, it has been reported that the degree of glycemic control among diabetic patients at the time of their index PCI did not have an impact on long-term outcomes after successful PCI.^59–62
There are currently two ongoing trials to address the role of glycemia and cardiovascular disease in patients with type II diabetes. The BARI 2D will determine whether aiming for a HbA1c target of < 7% by providing more insulin exogenously versus increasing insulin sensitivity will lead to a better cardiovascular disease outcome in type II diabetic patients with angiographically proven CAD. The Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial is the second trial which will determine whether by using all available drugs, aiming for a HbA1c target of < 6% versus 7.5% will lead to a better cardiovascular disease outcome in type 2 diabetic patients at increased risk for CAD or with early signs of CAD. The (Future Revascularization in Patients with Diabetes Mellitus Optimal Management of Multi-vessel Disease) FREEDOM trial will compare CABG with tight glycemic control with drug-eluting stent PCI and abciximab will address the question of PCI versus CABG using contemporary techniques.

Based on the above, it would seem prudent that diabetics should be managed using a team consisting of a diabetologist, cardiologist and a primary care physician who need to communicate better, work together more, and pursue common therapeutic goals. Cardiovascular disease risk factors: LDL cholesterol, hypertension, smoking, triglycerides and HDL cholesterol need to be aggressively treated. Beta-blockers, angiotensin-converting enzyme (ACE) inhibitors and angiotensin receptor blockers must be used appropriately. HbA1c should be lowered to < 7%.

Summary and Conclusions
Optimal PCI outcomes can be achieved with a strategy targeting the lesion site as well as the diseased diabetic vasculature. Gp IIb/IIIa antagonists with stenting equalize the outcomes in diabetics similar to non-diabetics. Drug-coated stents seem to be a promising new strategy to reduce the incidence of restenosis in diabetics. However, more long-term data is needed before the incremental efficacy of drug-coated stents in diabetics can be established. Diabetics derive an incremental benefit compared to non-diabetics with the use of brachytherapy. Lipid-lowering therapy has been proven to be of utmost importance in reducing coronary events after PCI, especially in diabetics. UKPDS revealed a trend towards reduction in coronary events with control of hyperglycemia in diabetics. BARI 2D, ACCORD and FREEDOM will shed light on this subject in the near future. At the present moment, PCI should be favored for focal lesions, larger vessels or where LIMA is not possible, re-operations, or patients with co-morbid conditions. In patients with LIMA being planned and those with associated valve disease, CABG should be favored for patients with diffuse disease, small vessels, type 1 insulin dependent diabetics, and good distal targets.

Address for correspondence: Umesh K. Arora, MD, The University of Texas Health Science Center at San Antonio, San Antonio, Texas. E-mail: umeen2111@aol.com