Glucose Lowering to Control Macrovascular Disease in Type 2 Diabetes: Title and subTitle BreakTreating the Wrong Surrogate End Point?

In the 1920s, the use of insulin to treat type 1 diabetes was lifesaving for children in diabetic ketoacidosis. Among the surviving patients with diabetes, the microvascular and macrovascular disease complications proved to be nonetheless devastating. The treatment of type 1 diabetes was revolutionized by the discovery that intensive glycemic control could prevent or delay the development of the microvascular complications of retinopathy, neuropathy, and nephropathy. Indeed, for patients with type 1 diabetes, aggressive insulin treatment also reduced the long-term risk of cardiovascular disease.¹

Therapeutic enthusiasm for intensive treatment expanded to include patients with type 2 diabetes, who typically have insulin resistance rather than the absence of insulin production characteristic of type 1 diabetes. Elevated glucose levels in patients with type 2 diabetes, like the high white blood cell counts in patients with bacterial pneumonia, are a consequence of insulin resistance together with inadequate compensatory hyperinsulinemia. Clinical trials of patients with type 2 diabetes demonstrated that improved glycemic control was associated with the prevention of microvascular complications. Numerous observational studies have demonstrated that high levels of glycated hemoglobin (HbA_1c) and glucose are predictors of cardiovascular disease in individuals with and without diabetes. The association even extends to glucose levels within the normal range. The relationship of glucose levels to cardiovascular disease mortality is especially strong in patients with established cardiovascular disease.² Whether reduction of cardiovascular disease risk would result from intensive glycemic control in patients with type 2 diabetes remains the unanswered hypothesis.

The Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial was designed to test precisely this hypothesis.³ All patients were randomized to receive either intensive therapy with a target HbA_1c level of 6.0% or standard therapy with a target level of 7.0% to 7.9%. The primary outcome was a composite of nonfatal myocardial infarction, stroke, or cardiovascular death. Standard therapy and especially intensive treatment required the use of multiple agents, which included metformin, glimepiride, repaglinide, rosiglitazone, acarbose, exenatide, insulin, and similar agents. These drugs, which all reduce serum glucose levels, were used at the discretion of the investigators and the patients. The primary hypothesis or model was that among patients with type 2 diabetes, glucose control as assessed by HbA_1c level, regardless of the specific drug, would reduce the risk of cardiovascular events. The fact that the trial was stopped 18 months early due to an increase in total mortality associated with intensive therapy raises questions about the drug therapies, the validity of this model, or both.

The failure in the ACCORD trial to provide evidence of a cardiovascular risk reduction associated with intensive glucose control has been replicated in other studies. In the Action in Diabetes and Vascular Disease: Preterax and Diamicron-MR Controlled Evaluation (ADVANCE) trial,⁴ 11 140 patients with type 2 diabetes were randomized to receive standard treatment or intensive glucose control that used gliclazide plus other drugs to achieve a target HbA_1c level of 6.5%. After a median of 5 years of follow-up, the investigators reported a reduced risk of the composite outcome of microvascular and macrovascular disease (hazard ratio, 0.90; 95% confidence interval, 0.82-0.98), primarily due to a risk reduction for nephropathy. For the composite outcome of macrovascular events, intensive treatment was associated with a nonsignificant risk reduction (hazard ratio, 0.94; 95% confidence interval, 0.84-1.06). Similarly, the Veterans Affairs Diabetes Trial, presented at the American Diabetes Association meeting in June 2008, failed to show a reduction in cardiovascular disease risk associated with intensive efforts to control glucose levels. Indeed, in the earlier Veterans Affairs Diabetes Feasibility Trial, intensive treatment was associated with a slightly higher risk of cardiovascular events.⁵

The paradox that blood glucose levels are strongly associated with the risk of cardiovascular disease but that glycemic control does not prevent cardiovascular disease may perhaps be explained by insulin resistance or hyperinsulinemia. Insulin levels also predict increased cardiovascular risk in large population-based studies. High fasting insulin concentrations were an independent predictor of ischemic heart disease events among 2103 Canadian men without diabetes.⁶ In the San Antonio Heart Study, homeostatic model assessment of insulin resistance (based on the product of fasting insulin with fasting glucose) was an independent predictor of incident cardiovascular events over 8 years of follow-up.⁷ In the Helsinki Policemen Study, 970 men free of diabetes or coronary heart disease at baseline underwent follow-up for 22 years; those with the highest levels of insulin area under the curve during oral glucose tolerance testing had the highest rates of coronary heart disease events and death.⁸ The correlation of insulin levels with cardiovascular disease raises the possibility that increases in levels of insulin, not glucose, may be etiologic in cardiovascular disease risk. Indeed, cardiovascular disease risk increases in the prediabetic state, before the development of overt hyperglycemia.

Aspects of insulin physiology suggest that it may play a direct role in vascular health. In addition to metabolic actions such as the suppression of hepatic gluconeogenesis and stimulation of myocyte glucose uptake, recent research has suggested that insulin may have protective, anti-inflammatory effects that would be predicted to retard atherosclerosis.⁹ On the other hand, insulin also has mitogenic, growth-promoting effects. Of particular relevance, insulin has been found to have mitogenic effects on vascular smooth muscle cells, increasing signaling in the mitogen-activated protein kinase pathway as well as promoting the proliferative effects of platelet-derived growth factor on these cells.¹⁰ Insulin also may increase atherosclerosis by increasing expression and activity of plasminogen activator inhibitor 1, decreasing fibrinolytic capability. If insulin levels are toxic to the cardiovascular system, then treatments designed to reduce insulin levels rather than glucose levels might be associated with a reduced risk of cardiovascular events in patients with type 2 diabetes.

The recent clinical trials with cardiovascular disease reduction as the primary end point used a mixture of antidiabetic therapies, some that increase the insulin level and others that decrease it. In part because multiple agents were required for intensive lowering of glucose levels and in part because various methods of lowering were assumed to be equivalent, the investigators avoided using multigroup trials that might have compared one agent or strategy with another agent or strategy of intensive therapy. Any effort to identify an agent that might be responsible for the increase in cardiovascular disease risk in these randomized clinical trials becomes a complex observational analysis.

Some prior clinical trial evidence supports the idea that therapies to lower insulin levels may be effective in preventing macrovascular disease. In the United Kingdom Prospective Diabetes Study (UKPDS), which used sulfonylureas, insulin, and metformin to lower glucose levels, metformin was the only agent associated with a reduction in macrovascular events.¹¹ Metformin treatment in overweight patients decreased myocardial infarction and stroke compared with conventional therapy and decreased all-cause mortality and total diabetes end points and stroke compared with insulin or sulfonylureas. With long-term follow-up, the reduction in the risk of myocardial infarction associated with metformin treatment was pronounced (relative risk, 0.67; 95% confidence interval, 0.59-0.89).¹² Of these 3 therapies, only metformin lowers insulin levels, while the others increase insulin levels.

If insulin levels increase the risk of cardiovascular disease, then the thiazolidinedione class of drugs, which reduce insulin levels, would in theory be expected to be especially effective. As agonists of the peroxisome proliferator-activated receptors, primarily the γ receptors, these drugs activate the transcription of a variety of genes, including those that affect metabolism of glucose and lipids. In the Prospective Pioglitazone Clinical Trial in Macrovascular Events (PROactive), pioglitazone was associated not only with a reduced risk of a composite outcome of major cardiovascular events but also with a much larger increase in the risk of heart failure.¹³ From the point of view of cardiovascular disease prevention, the off-target adverse effects of rosiglitazone—an increase not only of body weight, lipids, and fluid retention but also of fracture, heart failure, and perhaps myocardial infarction—also provide an imperfect or incomplete test of the insulin-lowering hypothesis.

If therapies to decrease insulin levels reduce macrovascular complications in patients with type 2 diabetes, then it is possible that some therapies to increase insulin levels may be associated with an increase in macrovascular complications. The balance of evidence suggests this is not the case, though with some exceptions. The University Group Diabetes Program (UGDP) found an excess of cardiac events with sulfonylurea therapy, possibly specific only to first-generation sulfonylureas, which can interfere with ischemic preconditioning.¹⁴ Among insulin-treated elderly participants in the Cardiovascular Health Study, those with the highest insulin levels, which reflected high insulin doses, had increased mortality over time.¹⁵

It is a challenge to disentangle the effects of the interrelated characteristics of glucose, insulin, and insulin resistance. Each may serve as a proxy for the other in epidemiologic studies of cardiovascular disease. In epidemiologic studies, moreover, the cardiovascular disease association may be stronger for glucose than for insulin in part because insulin levels are assayed with more measurement error than glucose levels. Today's therapies to decrease insulin levels do so by ameliorating insulin resistance. Thus, any benefits on cardiovascular disease risk may be due to insulin sensitization as well as—or instead of—lowering of ambient insulin concentrations.

Because several recent trials evaluating the strategy of lowering glucose levels have shown little or no benefit in terms of cardiovascular disease prevention in patients with type 2 diabetes, it may be appropriate to focus also on the aggressive control of insulin levels or insulin resistance rather than only on the aggressive control of glucose levels. Future trials of cardiovascular disease prevention in type 2 diabetes should test specific insulin-lowering agents or strategies rather than allowing multiple agents to be used with the goal of simply lowering HbA_1c levels. For the prevention of cardiovascular disease in patients with type 2 diabetes, intensive treatment of glucose levels may resemble aggressive efforts to reduce white blood cell counts in patients with bacterial pneumonia.