Little benefit to revascularization over intensive medical treatment in DM with CAD

Clinical Question

What is the best approach to the co-management of heart disease and diabetes mellitus?

Bottom Line

This complex study found little advantage to revascularization over intensive medical therapy in diabetic patients with coronary artery disease. Similarly, the authors found few differences between insulin sensitization and insulin provision, although metformin (known to reduce cardiovascular events) and thiazolidinedione (thought to increase them) were used in the same group and may have negated each other. Also, there were 12 combinations of any revascularization, percutaneous coronary intervention (PCI), or coronary artery bypass graft (CABG) versus medical therapy, insulin provision versus sensitization, and outcomes of mortality and cardiovascular events. Thus, it may have just been by chance that 1 of the 12 outcomes showed a benefit (the subgroup of patients whose physician thought they needed CABG, who underwent CABG, and were randomized to the insulin sensitization group). (LOE = 1b)


BARI 2D Study Group, Frye RL, August P, et al. A randomized trial of therapies for type 2 diabetes and coronary artery disease. N Engl J Med 2009;360(24):2503-2515.  [PMID:19502645]

Study Design

Randomized controlled trial (single-blinded)




Outpatient (any)


This large clinical trial used a Latin square, or 2 x 2 design, to attempt to answer 2 important questions simultaneously. A total of 2368 patients with type 2 diabetes and coronary artery disease who were candidates for elective PCI or CABG were enrolled in the study. Their cardiologist was asked to recommend a preferred intervention: either PCI or CABG. The 763 patients selected for CABG were then randomly assigned to undergo revascularization with CABG plus intensive medical therapy or intensive medical therapy alone. Similarly, the 1605 patients selected for PCI were randomized to undergo either PCI plus intensive medical therapy or intensive medical therapy alone. Then, each of these 4 groups was randomized to receive either insulin provision or insulin sensitization therapy for their diabetes, both with a goal of a hemoglobin A1C level of less than 7.0%. Patients in the insulin provision group were primarily treated with insulin and/or a sulfonylurea, while those in the insulin sensitization group were largely treated with metformin and/or a thiazolidinedione. The average age of participants was 62 years, 70% were men, 17% were black, and 13% were Hispanic (Am Heart J 2008;156:528-36). Patients were followed for a mean of 5.3 years. Groups were balanced and the analysis was by intention to treat. There was some crossover, of course, as in any such study: 13% assigned to medical therapy only were revascularized within 6 months; 42% by the end of the study. This compares with 95% of patients in the revascularization arms of the study who were revascularized within 6 months. Bare metal stents were used by approximately two thirds of patients, mostly in the earlier years of the patient recruitment period. There was no difference in all-cause mortality between patients receiving revascularization plus medical therapy versus those receiving intensive medical therapy only, or between patients receiving insulin sensitization or insulin provision. This was true whether they received PCI or CABG. There was also no difference between these groups regarding the likelihood of major cardiovascular events. The only exception was in the stratum whose physician thought they should consider CABG: patients who underwent CABG were more likely to remain free of a cardiovascular event than those receiving intensive medical therapy alone (77.6% vs 69.5%; P = .01; number needed to treat [NNT] = 12). However, that was only true for those assigned to the insulin sensitization group (18.7% vs 32.0%; P = .002; NNT = 8), not for those assigned to insulin provision (26.0% vs 29.0%; P = 0.58).