Implications of Contemporary Clinical Trials Cardiovascular Clinical Trials in Patients With Diabetes Mellitus Lessons From the Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) Study

Diabetes mellitus is a well-established risk factor for virtually all cardiovascular outcomes, with clinical trials and observational studies demonstrating greater mortality, more myocardial infarctions, and more episodes of heart failure in diabetic than in nondiabetic individuals.1–6 Moreover, diabetes mellitus is often associated with other cardiovascular risk factors, including hypertension and hyperlipidemia. The increased risk for adverse cardiovascular outcomes has made diabetic patients a particularly relevant target group for therapies that reduce cardiovascular risk because diabetic patients appear to benefit as much as or more than nondiabetic patients from successful cardiovascular therapies.7–9 Clinical trials that have assessed strategies to reduce cardiovascular risk in diabetic populations have generally focused on treatment of factors that have been linked to higher risk of cardiovascular disease, such as blood pressure, lipid levels, or albuminuria. Trials of therapies for these risk factors in exclusively diabetic populations benefit from higher overall event rates, which translate to increased power in clinical trials, so that therapies can be tested in fewer patients at lower cost. Although this approach may come at the expense of a broad indication for a specific therapy, the increasing worldwide prevalence of diabetes mellitus somewhat mitigates this concern and makes trials exclusively in diabetic populations financially more palatable for sponsors. Another reason to test therapies exclusively in a specific patient population is the finding of differential benefit in that population. Such a differential benefit in diabetic patients was observed in a subgroup analysis of the original Bypass Angioplasty Revascularization Investigation (BARI) trial, which randomized patients with multivessel coronary disease to either percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG).10 Although BARI showed no difference between those randomized to PCI or those randomized to CABG, CABG appeared superior in the subgroup of diabetic patients. Whether this differential benefit was a result of the diabetes mellitus per se or simply because these patients were at higher risk in general remains unclear. Observational data linking worse glycemic control to higher rates of cardiovascular events in diabetic patients have provided impetus for assessing the impact of therapies that improve glycemic control on cardiovascular risk. Although clinical trials have shown that therapies that improve glycemic control reduce the risk of microvascular disease, including retinopathy, nephropathy, and neuropathy,11 trials attempting to reduce macrovascular events have not proven successful; in the Action to Control Cardiovascular Risk in Diabetes (ACCORD),12 Action in Diabetes and Vascular Disease: Preterax and Diamicron Modified Release Controlled Evaluation (ADVANCE),13 and Veterans Affairs Diabetes Trial (VADT)14 studies, improved glycemic control did not reduce the rate of cardiovascular events and in some cases was even associated with increased risk. The BARI 2 Diabetes (BARI 2D) trial15 assessed the effect of both a cardiovascular intervention and a glycemic intervention on cardiovascular risk in a diabetic population. By utilizing a factorial design, BARI 2D was able to test efficiently and cost-effectively 2 separate hypotheses in the same population. Patients with diabetes mellitus and coronary disease were randomly assigned to either prompt revascularization with intensive medical therapy or intensive medical therapy alone and were also randomly assigned to either insulin-sensitization (including metformin or thiazolidinedione) or insulin-provision (insulin or sulfonylurea) therapy. The decision about the type of revascularization that would be performed (CABG or PCI) was made on clinical grounds before randomization to revascularization or intensive medical therapy. Importantly, this design did not allow BARI 2D to test the hypothesis that was generated by the subgroup analysis of the original BARI study because the decision to revascularize with PCI or CABG was clinically driven. Although BARI 2D showed no difference in the primary end point of death, myocardial infarction, or stroke between patients randomized to revascularization therapy plus intensive medical therapy versus intensive medical therapy alone, or those randomized to insulin-sensitization therapy versus insulin-provision therapy, this study illustrates a number of