The FREEDOM Trial: Revascularization in Diabetics with Multivessel Disease: A Population-Based Evaluation of Outcomes

The topic of revascularization in diabetics with multivessel disease is one which has perplexed cardiologists for years. In 1997, the BARI (Bypass Angioplasty Revascularization Investigation) trial compared balloon angioplasty with bypass surgery and tracked the results to 10 years, demonstrating a mortality benefit in coronary artery bypass graft (CABG) compared to balloon angioplasty in treated diabetics.1 Subsequent analysis was performed comparing bare metal stents (BMS) to CABG, and the ARTS-I (Arterial Revascularization Therapy Study) trial showed that the percutaneous coronary intervention (PCI) strategy required more frequent repeat revascularization, though there was not a significant difference in major adverse cardiovascular events.2

The advent of drug eluting stents prompted a re-examination of these results, and the FREEDOM (Future Revascularization Evaluation in Patients with Diabetes Mellitus: Optimal Management of Multivessel Disease) trial sought to put the question to rest. The FREEDOM trial was a prospective, randomized, multicenter superiority trial that compared multivessel PCI to CABG in diabetic patients on optimal medical therapy.3 The National Heart, Lung and Blood Institute sponsored trial demonstrated CABG was superior to PCI with first generation drug-eluting stents. Outcomes favoring CABG were driven by a reduction in all-cause mortality and non-fatal myocardial infarction (MI) while complicated by increased rates of stroke. The question was seemingly answered, with significantly higher major adverse cardiac and cerebrovascular event (MACCE) in the PCI group (26.6% vs 18.7%) and a number needed to treat (NNT) of 12.6 favoring CABG. These results seemed to confirm those which had preceded them, including the SYNTAX (TAXUS Drug-Eluting Stent Versus Coronary Artery Bypass Surgery for the Treatment of Narrowed Arteries) trial, of which the medically treated diabetic subset showed very similar rates of MACCE (26.0 for PCI vs 14.2% for CABG).4

Given the overwhelming evidence in well executed trials, Dr. Ramanathan and colleagues sought to explore the real world practice patterns in British Columbia using registry data.5 They used the background of the FREEDOM data to look at revascularization strategy trends in diabetic patients with multivessel disease from 2007 to 2014, which showed that despite publication of the FREEDOM data in 2012, clinical practice had not changed as significantly as might have been expected.

Of 51,203 revascularization cases, they identified 4,819 procedures on diabetics with multi vessel coronary disease consistent with the FREEDOM trial exclusion data. Of these, the dominant revascularization strategy was PCI (60% PCI vs 40% CABG). Using their databases, and ICD-10 codes, as well as data from vital statistics, they looked for MACCE. Once again, they found that the real life registry results mirrored the randomized trial results. The composite endpoint favored CABG (adjusted and non-adjusted) with a short term relative risk reduction of 39% (30 day) and long term risk reduction of 36% (31 day to 5 years). Even among secondary outcomes, the investigators found that long term secondary outcomes, including death, MI, stroke, and revascularization rate were all more significant in the PCI group than the CABG group. While the short term stroke rate was higher among those undergoing CABG, in this study the long term results demonstrated lower stroke in the CABG group compared to PCI. This is of particular interest as previous randomized trials have consistently shown numerically higher stroke rates in the CABG cohorts.4

Nearly two-thirds of the procedures were performed in patients considered to be stabilized acute coronary syndrome and nearly two-thirds of these revascularizations were with PCI (65% PCI vs 35% CABG). Given the lower mortality in the CABG cohort, the authors conclude that CABG may be worthy of consideration even in the stabilized acute coronary syndrome (ACS) patient population.

Also of great value in this data, in our opinion, is in looking at the baseline patient characteristics, as they tell us what is happening to patients in actual practice. PCI patients were older, more likely to be female, have significantly more pulmonary disease, were more likely to have stabilized ACS vs. stable ischemic disease, were more likely to require urgent revascularization, and were more likely to have an ejection fraction <50%. All of these portend higher risk patients, and a greater likelihood of future MACCE.

Furthermore, the median time from angiography to PCI was less than 1 day. The question is whether this means that the procedures are being done ad hoc, or because doctors know, either through consultation with a cardiothoracic surgeon or through prior knowledge of surgical preferences, whether a patient would be considered an appropriate candidate for surgery.

As seen in this study, those undergoing CABG are more likely to have three vessel disease or proximal left anterior descending (LAD) disease, which anatomically is generally considered standard clinical practice. In the FREEDOM trial, if we recall, 83% of patients had three vessel disease, while in this study, 2049 of 4819 procedures were categorized as three vessel. Furthermore, 64% of the CABG population had three vessel disease, while only 28% of PCI patients in this study had three vessel disease. Of those undergoing CABG, 45% had proximal LAD involvement, and only 23% of the PCI patients had proximal LAD disease (left main disease was excluded), suggesting CABG was preferred in diabetic patients with multivessel disease and proximal LAD involvement. What is not as clearly defined is the population with two-vessel disease and those with less complex multivessel disease. We have learned from the SYNTAX trial that complexity of anatomy predicts mortality and that PCI is a reasonable alternative to CABG in less complex multivessel disease (SYNTAX ≤22) as mortality benefit is comparable among CABG and PCI. Subgroup analyses to see the interrelation between some of these lesion characteristics and the presentation as ACS versus stable ischemic heart disease (SIHD) would be interesting, and may shed further light on the patient characteristics of this study.

The authors themselves recognize the limitations of the data, particularly on the complexity of the anatomy and completeness of revascularization on this observational registry data. The anatomical complexity is normally paramount to any such decision making. They also recognize the potential survival bias in the CABG cohort. Another important limitation is the length of follow up. The long term durability of CABG, particularly with non-arterial grafts, is of important consideration. While 5 years patency for CABG is well established, the advantages begin to slip away at 10-15 years. Furthermore, this study observes the time period from 2007-2014, when second generation drug eluting stents were implemented, and does not give data on which type of stents were placed.

As new data emerges and additional methods for risk stratification are established, we will almost assuredly approach the question of what to do in diabetic patients with multivessel disease, whether with SIHD or with stabilized ACS. However, the most impressive question may not be what, but who. Who are the patients who may be too high risk for CABG, or who may not be anatomically suitable for PCI? And of course, the who is also largely dependent on patient preferences. The patients should certainly be presented with this data, including short and long term outcomes, and we believe that a discussion with a cardiothoracic surgeon and implementation of the heart team approach should take place. If suitable, the overwhelming data for most multivessel disease, particularly if a left internal mammary artery graft can be utilized, is for CABG. The question arises in those diabetic patients whose multivessel disease is less complex or those who are not good surgical candidates. Thus, the decision for revascularization strategy is based on a continuum of factors and the heart team discussion is a vital component of the decision making process. This is where further research in the appropriate patient characteristics will serve our field well.


  1. BARI Investigators. Comparison of coronary bypass surgery with angioplasty in patients with multivessel disease. N Engl J Med 1996;335:217-25.
  2. Abizaid A, Costa MA, Centemero M, et al. Clinical and economic impact of diabetes mellitus on percutaneous and surgical treatment of multivessel coronary disease patients: insights from the Arterial Revascularization Therapy Study (ARTS) trial. Circulation. 2001;104:533-8.
  3. Farkouh ME, Domanski M, Sleeper LA, et al. Strategies for multivessel revascularization in diabetic patients.N Engl J Med 2012;367:2375-84.
  4. Banning AP, Westaby S, Morice MC, et al. Diabetic and nondiabetic patients with left main and/or 3-vessel coronary artery disease: comparison of outcomes with cardiac surgery and paclitaxel-eluting stents. J Am Coll Cardiol 2010;55:1067-75.
  5. Ramanathan K, Abel J, Park JE, et al. Revascularization in patients with diabetes and multivessel coronary artery disease: A population based evaluation of outcomes. American Heart Association 2015.

Clinical Topics: Acute Coronary Syndromes, Cardiac Surgery, Diabetes and Cardiometabolic Disease, Invasive Cardiovascular Angiography and Intervention, Aortic Surgery, Interventions and ACS, Interventions and Coronary Artery Disease, Interventions and Imaging, Angiography, Nuclear Imaging

Keywords: Acute Coronary Syndrome, Angiography, Angioplasty, Balloon, Coronary, Coronary Artery Bypass, Coronary Artery Disease, Decision Making, Diabetes Mellitus, Drug-Eluting Stents, International Classification of Diseases, Lung Diseases, Mammary Arteries, Myocardial Infarction, Percutaneous Coronary Intervention, Prospective Studies, Registries, Research Personnel, Risk, Risk Reduction Behavior, Stroke, Surgeons, Taxus, Vital Statistics, Metabolic Syndrome X

< Back to Listings