Review of the BARI 2D trial
Revascularization versus medical therapy in patients with type 2 diabetes
N Engl J Med 2009;360:2503-2515
Background: Type 2 diabetes increases the risk of cardiovascular events and death. Previous trials comparing revascularization versus medical therapy included patients with diabetes, however, a large-scale trial specifically focusing on patients with type 2 diabetes was lacking.
The Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) trial sought to assess the optimal treatment strategy for patients with type 2 diabetes and stable coronary artery disease.
Patients: Eligible patients had type 2 diabetes and stable coronary artery disease. Coronary artery disease was defined as a stenosis in a major coronary artery of 50% or more and a positive stress test or 70% or more and classic angina. Patients had to be candidates for percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG) without further specification.
Patients were excluded if they had left main disease, prior PCI or CABG within 12 months, class III or IV heart failure, hepatic dysfunction, creatinine> 2 mg/dL or glycated hemoglobin> 13%.
Baseline characteristics: The trial randomized 2,368 patients – 1,176 randomized to the revascularization arm and 1,192 to the medical therapy arm. Among the 1,176 patients in the revascularization arm, 32% were planned to undergo CABG and 68% planned to undergo PCI.
The average age of patients was 62 years and 70% were men. The mean glycated hemoglobin was 7.7% and the mean duration of diabetes was 10.4 years. Approximately 32% had prior myocardial infarction, 7% had congestive heart failure and 24% had peripheral artery disease. Approximately 18% had no angina or angina equivalent. Angina class within 6 weeks was 1-2 in 43% of the patients and 3-4 in 9%.
The mean left ventricular ejection fraction was 57%. Approximately 31% had three-vessel disease and 13% had proximal left anterior descending artery disease.
Baseline characteristics were well balanced between the revascularization arm and the medical therapy alone arm. However, patients who were in the CABG stratum had more three-vessel disease (52% vs 20%) and more proximal left anterior descending artery disease (19% vs 10%).
Procedures: The trial was a 2 x 2 factorial design and patients were randomly assigned to two treatment strategies. The first was randomization to revascularization or medical therapy. The second was randomization to insulin-sensitization therapy or insulin-provision therapy. Randomization was stratified based on the method of revascularization (PCI vs CABG) which was determined by the treating physician.
In this review, we will focus on the first strategy of revascularization vs medical therapy.
For patients randomized to the revascularization arm, the procedure was to be performed within 4 weeks after randomization. Patients in the medical arm could receive revascularization on follow up for any of the following: Progression of angina, acute coronary syndrome or severe ischemia.
Patients were seen monthly for the first 6 months and every 3 months thereafter.
Endpoints: The primary endpoint was death from any cause. Secondary end point was a composite of death, myocardial infarction, or stroke.
Analysis was performed based on the intention-to-treat principle. The original sample size of 2,800 patients was not met, and therefore, the average follow up time was increased by 1.5 years to become 5.3 years. Using the new follow up duration, the study had 88% power to detect a 33% relative risk reduction of death (from 14.0% to 9.8%), and a 95% power to detect a 25% relative risk reduction in the secondary composite endpoint (from 24.0% to 18.0%).
Results: Among the patients randomized to the revascularization arm, 95.4% underwent revascularization at 6 months compared to 13.0% of the patients randomized to the medical arm. At 5-years, 42.1% of the patients randomized to the medical arm had undergone revascularization. Among patients who underwent PCI in the revascularization arm, procedures were attempted on average of 1.5 lesions and 56.0% received a bare metal stent. Among patients who underwent CABG in the revascularization arm, 94.2% received an internal mammary artery graft and the mean number of distal anastomoses was 3.0.
The average follow up time was 5.3 years.
There was no significant difference in the primary outcome of all-cause death. Survival was 88.3% in the revascularization arm and 87.8% in the medical arm (difference: 0.5%; 95% CI: −2.0 - 3.1; p=0.97). There was also no significant difference for the secondary composite endpoint. Freedom from events for the secondary endpoint was 77.2% in the revascularization arm and 75.9% in the medical arm (difference: 1.3%; 95% CI, −2.2 - 4.9; p=0.70).
Survival was not significantly different between both treatment strategies in the CABG stratum (86.4% with revascularization vs 83.6% with medical therapy; p= 0.33). However, patients in the CABG stratum had more freedom from the secondary composite endpoint (77.6% vs 69.5%; p= 0.01).
In the PCI stratum, revascularization did not improve survival (89.2% with revascularization vs. 89.8% with medical therapy; p= 0.48) or freedom from the secondary composite endpoint (77.0% with revascularization vs 78.9% with medical therapy; p= 0.15).
Conclusion: In patients with type 2 diabetes and stable coronary artery disease, revascularization compared to medical therapy did not improve the primary outcome of all-cause death, or the composite secondary outcome of death, myocardial infarction or stroke over an average follow up time of 5.3 years.
The observed benefit of revascularization within the CABG stratum should be viewed as hypothesis-generating rather than conclusive evidence that CABG is superior to PCI in this patient population.
One potential limitation of this trial is that the authors included patients who were candidates for either PCI or CABG without providing enough details on what makes someone not a candidate. This lack of clarity limits physicians' ability to fully understand which patients would have been suitable for inclusion.
I would say that some ambiguity about “candidate for either CABG or PCI” improves the external validity of this result, since on a practical level we actually know who these pts are. The Cath doc has rendered an opinion. You have a surgeon review the films. You’re not dealing with CTO’s or very high Syntax score anatomy. And you’re not dealing with poor overall surgical candidates. If both say they would take the case, then that’s a BARI-2D pt. If one of them says they wouldn’t take the case, then you also have your answer.
Btw this study and Freedom, it is unfortunate there has not been a 3 arm study of PCI vs CABG vs med therapy in this population.
You correctly point out the important practical limitation to this study, “One potential limitation of this trial is that the authors included patients who were candidates for either PCI or CABG without providing enough details on what makes someone not a candidate. This lack of clarity limits physicians' ability to fully understand which patients would have been suitable for inclusion.”
Why not address this in the form of a letter to the editor of NEJM inviting the authors to respond.
On a broader level, journals might provide a section for authors of accepted articles to provide an addendum, even many years later, providing answers to important questions not asked by the reviewers and not answered in the original publication.