Statistics from Altmetric.com
Patients with diabetes mellitus have increased rates of coronary artery disease (CAD), myocardial infarction (MI), heart failure and death. Importantly, approximately 25% of patients undergoing coronary artery revascularisation have diabetes.1 Recently, the Future Revascularisation Evaluation in Patients with Diabetes Mellitus: Optimal Management of Multivessel Disease trial found coronary artery bypass graft (CABG) surgery superior to percutaneous coronary intervention (PCI) in significantly reducing rates of death and MI.2 It has been suggested that this study settles the controversy over the comparative effectiveness of CABG versus PCI for patients with diabetes and stable multivessel CAD.3 ,4 However, the optimal revascularisation strategy in this patient subgroup has been debated for years because of conflicting reports.
Early results comparing CABG with PCI in patients with diabetes came from randomised clinical trial subgroup analyses.5 Only one reported a significant survival advantage for CABG,6 but small sample size and short follow-up duration were limitations of many trials. In the Bypass Angioplasty Revascularisation Investigation (BARI) trial,6 the 353 patient subgroup with diabetes had a lower 5-year mortality rate with CABG than with balloon angioplasty (19.4% vs 34.5%, p=0.0024). In 1995, the US National Heart, Lung, and Blood Institute issued a clinical alert emphasising that finding, and supporting CABG as the revascularisation strategy of choice in patients with diabetes.7 After 10 years of follow-up, survival remained superior with CABG (57.8% vs 45.5%, p=0.025).8 Repeat 10-year revascularisation rates were 20.3% with CABG and 76.6% with PCI (p<0.001). A collaborative analysis performed on 1233 patients with diabetes included in 10 randomised trials also demonstrated a lower 5-year mortality rate with CABG compared with balloon angioplasty or bare metal stent implantation (12.3% vs 20%, HR 0.70, 95% CI 0.56 to 0.87).5
More recently, the BARI 2 Diabetes trial enrolled 2368 patients with diabetes and found no difference in 5-year rates of death or major cardiovascular events between optimal medical therapy and coronary artery revascularisation.9 Because patients were stratified to CABG or PCI cohorts by anatomical complexity of disease before randomisation, the revascularisation strategies could not be directly compared. Nevertheless, freedom from major cardiovascular events was better with revascularisation in the CABG stratum (77.6% vs 69.5%, p=0.01), but not in the PCI stratum. However, survival was not significantly different with CABG compared with medical therapy (86.4% vs 83.6%, p=0.33) or with PCI compared with medical therapy (89.2% vs 89.8%, p=0.48).
The SYNergy between PCI with TAXus and cardiac surgery trial has published 3-year results in the 452 patient subgroup with diabetes.10 The composite rate of death, MI and stroke was not different between drug-eluting stents (DES) and CABG (16.3% vs 14.0%, p=0.527), but repeat revascularisation rates were higher with PCI (28.0% vs 12.9%, p<0.001).
Patients in registries receive PCI or CABG by choice, rather than according to random assignment. Patients eligible but not randomised in the BARI Trial were included in the BARI Registry.11 In contrast with the randomised trial results, survival was not significantly different in the diabetic subgroup that included 182 patients treated with PCI and 117 patients treated with CABG (14.4% vs 14.9%). A pooled analysis of 8818 patients with diabetes from seven registries with follow-up ranging from 5–12 years showed a long-term mortality of 27.8% with PCI compared with 26.3% with CABG.12
The Coronary Artery in Revascularisation Diabetes trial was the first prospective randomised trial comparing revascularisation strategies in patients with diabetes.13 A total of 510 patients with multivessel or complex CAD were enrolled and followed for 5.1 years. The combined rate of death, MI and stroke was 26.6% with PCI versus 20.5% with CABG (HR 1.34; 95% CI 0.94 to 1.93, p=0.11). Mortality rates were similar (14% vs 12.6%), but non-fatal MI rates (14% vs 6.3%) and repeat revascularisation rates were higher with PCI, whereas non-fatal stroke rates were higher with CABG (4.3% vs 3.1%).
The Future Revascularisation Evaluation in Patients with Diabetes Mellitus: Optimal Management of Multivessel Disease Trial randomised 1900 patients from 140 international centres and followed them for a median of 3.8 years.2 Mean left ventricular ejection fraction was 66%, and greater than 40% in 98% of patients; three-vessel CAD was present in 83%. The estimated composite outcome of death, MI or stroke at 5 years was 26.6% with PCI using DES versus 18.7% with CABG (p=0.005) (figure 1). Death (10.9% vs 16.3%, p=0.049) and MI (6.0% vs 13.9%, p<0.001) rates were lower with CABG, whereas stroke rate was higher (5.2% vs 2.4%, p=0.03). The mortality rate only became statistically different after 4 years (figure 1). It should be noted that the original protocol was designed to enrol 2400 patients to ensure a power of 85% to detect a relative reduction of 18–23% in 4-year event rates of the composite outcome. Two subsequent protocol amendments changed the target sample size to 1900 patients with an estimated power of 80% to detect a relative reduction of 27% in the 4-year event rates in the two study groups. Since only a quarter of patients were followed for more than 4 years, longer and more complete follow-up is needed to accurately quantify the study results.
Clinical practice guidelines
The 2010 European Society of Cardiology and the European Association for Cardio-Thoracic Surgery guidelines on myocardial revascularisation offer three specific recommendations for diabetic patients with stable CAD.14 First, revascularisation is indicated for patients with extensive CAD to improve freedom from major cardiovascular events (Class I, Level A). Second, DES are recommended for PCI to reduce restenosis and repeat target vessel revascularisation rates (Class I, Level A). Finally, CABG should be considered, rather than PCI, when the extent of CAD justifies a surgical approach (especially multivessel disease) and the patient's risk profile is acceptable (Class IIa, Level B). The 2011 American College of Cardiology Foundation and American Heart Association CABG and PCI guidelines state that CABG is probably recommended in preference to PCI to improve survival in patients with multivessel CAD and diabetes, particularly if a left internal mammary (LIMA) graft can be anastomosed to the left anterior descending (LAD) artery (Class IIa, Level B).15
The evidence base and the clinical guidelines support a survival advantage for CABG in patients with diabetes and multivessel CAD at the population level. One limitation of the randomised clinical trial results is that patients had to be technical and clinical candidates for CABG and PCI and were therefore carefully selected, so generalisability can be called into question. Conversely, observational studies that have not consistently shown a survival advantage with CABG over PCI in patients with diabetes were confounded by treatment selection bias, suggesting that clinical decision-making has a major impact on prognosis after coronary revascularisation. On an individual patient level, reasons for not selecting CABG in all patients with diabetes include absence of left main or proximal LAD disease, poor surgical targets because of diffuse or distal disease, concern about developing saphenous vein bypass graft disease in patients with expected long-term survival, or comorbid disease that increases the risk for perioperative complications. For instance, a cardiologist might be more inclined to support PCI for a patient with younger age, ≤four significant simple stenoses, normal left ventricular function, or significant comorbid disease (severe cerebral, renal, pulmonary, hepatic disease; age, frailty). Conversely, CABG would probably be a better recommendation for the patient with a greater atherosclerotic burden (diffuse disease or >four significant stenoses), more complex disease (chronic total occlusions, long calcified stenoses, bifurcation stenoses not easily treated with PCI) or left ventricular ejection fraction <40%.
Informed decision-making also requires consideration of patient preference. PCI will be attractive to some as the initial treatment strategy because it is less invasive and avoids perioperative mortality risk, surgical complications (reoperation for bleeding, pneumonia, wound infections, stroke, pleural effusions, atrial fibrillation) and delayed recovery. CABG still can be performed later in the disease process as a complementary, not competitive, revascularisation strategy, hopefully only once in a patient's life. Others will choose to accept the surgical risks with the expectation that there is a statistical survival benefit with CABG and a lower risk for repeat revascularisation procedures.
Surgical skill and the surgical revascularisation strategy may be other important variables. The likely explanation for the survival benefit associated with CABG is continued patency of a LIMA graft to the mid-LAD, treating the initial culprit lesion and potential new proximal disease. Stent implantation only treats the culprit lesion. In contrast, saphenous vein grafts fail at a consistent rate, with as many as 50% diseased or occluded within 10 years. Repeat revascularisation, if indicated, is challenging. Repeat CABG provides less improvement in quality of life and higher morbidity and mortality risks than the first surgery. Compared with native artery PCI, saphenous vein graft PCI has higher no-reflow and periprocedural MI rates, and higher rates of restenosis, disease progression and stent thrombosis. Also, CABG without a LIMA-LAD bypass graft may not offer prognostic superiority over PCI with DES implantation. CABG only with arterial grafts probably is superior to the standard operation with a LIMA-LAD graft and saphenous vein grafts to the other targets, but few surgeons perform that procedure. Indeed, a hybrid procedure might ultimately be the best revascularisation strategy with a LIMA-LAD bypass graft to improve prognosis and DES stent implantation for other stenoses to reduce ischaemia and avoid the potential long-term complication of saphenous vein bypass graft disease.
In summary, the best revascularisation strategy for a patient with diabetes and multivessel CAD is based on the complexity and extent of CAD, the magnitude of left ventricular systolic dysfunction, physician judgment, patient preference and consultation between the interventional cardiologist and the cardiac surgeon. Included in this discussion should be the potential survival benefit associated with implantation of a LIMA-LAD bypass graft; the increased risk of early stroke, other surgical complications, and the potential late complication of saphenous vein bypass graft disease with CABG; and the increased risk of MI and repeat revascularisation procedures after PCI. Choosing the best revascularisation strategy for an individual patient is a complex decision best decided by the heart team, not a randomised trial or an observational registry report.14 ,15
Competing interests None.
Provenance and peer review Commissioned; internally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.