Evidence of the Month

Commentaries on both new and classic studies of importance for the treatment of diabetes are posted here monthly.

BARI 2D: Comparison of Therapies for Ischaemic Heart Disease and Type 2 Diabetes

Bernard Charbonnel Endocrinology Department University of Nantes Nantes, France	Click here to download PowerPoint slides of this commentary.
	Click here for the programme evaluation and post-test.

Background
The Bypass Angioplasty Revascularization Investigation (BARI) has been best known for its research on cardiovascular protection in individuals with ischaemic heart disease. The BARI team reported in 1996 that coronary revascularization methods—percutaneous balloon angioplasty versus coronary-artery bypass grafting (CABG)—for patients with multivessel disease achieved similar 5-year survival results (BARI Investigators, 1996). However, for the BARI patients who also had Type 2 diabetes, survival was significantly better after bypass surgery than after angioplasty. In the newly reported BARI 2 Diabetes (BARI 2D) trial, the investigators studied patients with diabetes and heart disease exclusively, this time asking whether prompt revascularization, be it percutaneous or surgical, offers better cardioprotection than intensive medical therapy for risk factors. In addition, BARI 2D asked a timely and relevant question about glucose control strategies: is therapy that improves insulin resistance more cardioprotective than therapy that improves insulin supply?

Evidence from clinical comparisons of these glucose control strategies has been scarce and conflicting with respect to cardiovascular impact. The United Kingdom Prospective Diabetes Study (UKPDS) 34, which compared standard (dietary) versus more intensive pharmacological (sulfonylurea, insulin, or metformin in overweight diabetic patients) glucose therapy included a secondary analysis of metformin versus sulfonylurea or insulin. The metformin-treated patients did significantly better on clinical outcomes including diabetes-related and all-cause mortality over a decade, at least when metformin was used alone and not in combination (UKPDS 34, 1998). An analysis of 3101 patients with diabetes and heart disease in the Sibrafiban Versus Aspirin to Yield Maximum Protection From Ischemic Heart Events Post-acute Coronary Syndromes (SYMPHONY) trial registries found that major clinical events were more than twice as frequent with insulin-providing strategies than with insulin-sensitizing strategies (McGuire et al, 2004). By contrast, analysis of 2395 patients with diabetes and heart disease in the Bezafibrate Infarction Prevention study registry found higher all-cause mortality associated with metformin use than with diet or sulfonylurea (Fisman et al, 1999). Sensitization therapy using thiazolidinediones has brought a lot of controversies (Nissen, Wolski, 2007), even though the outcome trials, PROspective pioglitAzone Clinical Trial In macroVascular Events (PROactive) and Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes (RECORD), are overall reassuring (Dormandy et al, 2009; Home et al, 2009). The BARI 2D trial was launched by the US National Institutes of Health with the specific goal of clarifying the long-term cardiovascular impact of glucose control strategies in high-risk patients.

Methods and Key Results
From 2001 to 2005, BARI 2D randomized 2368 participants with Type 2 diabetes and stable angiographically documented coronary ischaemia, some with multivessel disease. The treating physicians decided on CABG or percutaneous coronary intervention (PCI) as the most appropriate revascularization option for each patient. From the resulting 2 groups, patients were assigned to 2 treatment protocols in a 2 x 2 factorial design.

The first strategy consisted of either prompt (within 4 weeks) revascularization via the selected method or intensive medical therapy. Revascularization in the medical group took place only if needed for worsening ischaemia during the trial. For the second strategy, patients within each stratum—CABG, PCI, and their corresponding medical arms—received either insulin-sensitizing therapy (mainly with metformin and rosiglitazone) or insulin-providing therapy (sulfonylurea and insulin). Target values for all patients were HbA_1c <7 %, low-density lipoprotein cholesterol <100 mg/dl (2.6 mmol/l), and blood pressure ≤130/80 mmHg, and all patients were counselled about weight, exercise, and smoking cessation.

The primary end points of BARI 2D were all-cause mortality and a composite of major cardiovascular events: death, myocardial infarction, or stroke. After an average follow-up of 5.3 years, survival rates did not differ significantly between the participants who had revascularization (88.3%) and those who received medical therapy alone (87.8%); nor between the participants who received insulin-sensitizing therapy (88.2%) and those who received insulin-providing therapy (87.9%). Similarly, the rates of freedom from major cardiovascular events did not differ significantly among the groups: 77.2% with revascularization versus 75.9% with medical therapy; 77.7% with insulin sensitization versus 75.4% with insulin provision.

The only significant difference emerged when the CABG and PCI strata were analyzed separately: in the CABG stratum, the rate of major cardiovascular events was significantly lower after prompt revascularization (22.4%) than with medical therapy (30.5%) (P=0.01); the lowest event rate, 17.6%, was seen for CABG patients who also received insulin-sensitizing therapy. In addition, insulin sensitization was associated with significantly more individuals achieving targets for HbA_1c, low-density lipoprotein cholesterol, and blood pressure.

Episodes of severe hypoglycaemia during BARI 2D were significantly more common with insulin-providing therapy (9.2%) than with insulin-sensitizing therapy (5.9%) (P=0.003). Peripheral pitting edema was significantly more frequent with insulin sensitization (P=0.02). There were no significant differences between the groups in rates of congestive heart failure.

Clinical Implications
The BARI 2D results are reassuring in showing that the 2 main strategies for diabetes therapy have, overall, nearly equal effects with respect to long-term cardiovascular risk, whether they are used in conjunction with coronary revascularization or with intensive medical treatment. Favouring an insulin-sensitizing strategy were findings linking it to fewer hypoglycaemic episodes and greater achievement of HbA_1c and other treatment targets. Insulin sensitization was also associated with lower plasma insulin levels, less weight gain, and higher levels of high-density lipoprotein cholesterol than were seen with insulin-providing therapy.

Substrata analyses showed that the combination of CABG and insulin-sensitizing therapy was associated with significantly greater cardioprotection than any other strategy. These results argue that diabetic patients with more severe heart disease may be best served by prompt CABG along with insulin-sensitizing therapy, rather than by treatment with medication alone, especially sulfonylurea/insulin medication. On the other hand, for patients with milder heart disease who are candidates for percutaneous intervention, medication alone may be a suitable and safe first option, given that prompt PCI showed no protective advantage. To be sure, about 42% of patients assigned to medical therapy in BARI 2D required revascularization later in the course of the trial. Nevertheless, their overall outcomes, including cardiovascular status, did not appear to suffer from the delay. It should be noted that the CABG/PCI substrata analyses were not prespecified in the BARI 2D design. However, the findings concerning superior clinical benefits of CABG confirm those of the original BARI trial.

The BARI 2D authors point out that differences in outcomes for patients taking rosiglitazone, metformin, or both for insulin sensitization could not be determined. Further, the combination of insulin-sensitizing and insulin-providing therapies did not appear to result in adverse outcomes. This has been a concern raised by other trials; in the UKPDS 34, for example, the addition of metformin to sulfonylurea therapy was associated with a significantly increased risk of diabetes-related death compared with sulfonylurea treatment alone (UKPDS 34, 1998).

Certainly other influences, including economic pressures, have bearing on treatment choices for diabetic patients with heart disease. The BARI 2D finding that prompt revascularization offers no incremental benefit for most such patients may help sway clinical support toward medical therapy for first-line cardioprotection—a more economic approach than revascularization. Research should continue to mount similar comparisons as newer medications become available for glucose control and the reduction of other risk factors.