The intensive glucose-lowering arm of the ACCORD trial was just shut down because of an increased number of deaths in that portion of the study. That finding has created a great deal of confusion on the long-held belief that aggressive treatment of blood sugar should protect patients from vascular events and death. That belief, like many that prevailed in the practice of medicine, made perfect sense. Virtually, all of the bad things that happen to diabetics are the result of vascular disease. Blindness, kidney failure, and nerve injury are the result of small vessel disease. Heart attack, stroke, and amputation are the result of large vessel disease. For a long time, we have known that lower sugar reduces the damage to small vessels. Very aggressive care should, therefore, lower heart attacks and strokes due to large vessel disease.

Glucose lowering, per se, does not prevent heart attack, stroke, or death. I think the ACCORD trial did establish that fact, but this finding is now commonplace in vascular medicine. Let not your heart be troubled. There is a place for aggressive glucose control in optimal medical therapy for diseases of the large vessels. We have learned that certain drugs lower the risk factor but do not protect the patient from events or death. ACCORD is no different. Treating hypertension prevents heart attack and stroke. Alpha blockers lower blood pressure. Alpha blockers are inferior to other treatments in preventing events. Rosiglitazone lowers blood sugar. Rosiglitazone is associated with increased cardiovascular events. The very first trial that looked at glucose lowering using tolbutamide was shut down early because of an increased death rate. For decades, estrogen was prescribed for its cardioprotective effects. That practice ended when a controlled clinical trial showed that estrogens combined with progesterone were associated with increased numbers of heart attacks. We have known for decades that higher HDL lowers vascular risk, yet a clinical trial for a medication that dramatically raised HDL had to be shut down because of increased events. In fact, our whole system functions as if bypass surgery and angioplasty are the ultimate answer in heart attack prevention. These treatments do nothing to prevent heart attack in the stable patient. There is, therefore, a long list of treatments that make sense. But when subjected to clinical trials, they offer no benefit or actually cause harm.

The ACCORD trial was not a vascular trial but a sugar trial. The trial was designed to test the hypothesis that lowering the sugar, taking precedence above all other considerations would lower events. The interventions section from the ACCORD protocol states,

“Both the intensive and the standard therapy groups will utilize all currently available glucose-lowering therapies. The two treatment groups will have different glycemic targets and will have different thresholds of glycemic control at which therapeutic changes will be considered.

To achieve these glycemic targets, participants will require self-management education and dietary and lifestyle interventions, as well as pharmacologic therapy. They will also require different drug choices and treatment intensities. For example, within 6 months of randomization, most intensive group participants will likely be on 3 or more injections of insulin a day in addition to 2 or 3 oral agents. Conversely, standard therapy participants are less likely to be on insulin, will be on less than or equal to 2 injections per day if insulin is used, and will be taking fewer oral agents. Moreover, the frequency with which self-management behavior is applied and participants are contacted will vary between the two levels of glycemic control.”

Later, in the Dietary and Lifestyle Interventions section, there is the following guidance for investigators:

“All participants will be provided with the same dietary and lifestyle recommendations to optimize their glucose control. These will include advice that blood glucose control may be more critical than weight control in reducing the risk of complications of diabetes.”

The investigators could use Metformin, thiazolidinediones (TZDs), insulins, sulfonylureas, exenatide, and acarbose.

There have been serious questions raised in the past about the adverse effects of certain drugs for diabetes on vascular events and survival. These questions have been raised for rosiglitazone and sulfonylureas. Diabetes is, itself, a weight-related illness. Most type 2 diabetics have blood pressure and cholesterol problems that are both made worse by increased weight. All diabetic treatments except for Metformin, exenatide, and acarbose are associated with weight gain. If you compare twice daily insulin with Metformin and a single long-acting insulin shot, the Metformin-insulin people gain no weight and the twice daily insulin patients gain an average of 10 pounds in one year. This is no small matter in a weight-driven illness. The ACCORD trial allowed a number of medical interventions that would be expected to produce weight gain.

On the very same day that the intensive glucose-lowering arm of ACCORD was shut down, the Steno-2 investigators reported their 13-year follow-up for aggressive treatment of high-risk type 2 diabetic patients. They found that early intensive intervention with multiple drug combinations and behavior modification leads to reduced rates of death and cardiovascular disorders. There was a 20% absolute risk reduction in the primary endpoint of all-cause mortality. There was also a 13% absolute risk reduction for cardiovascular death. In Steno-2, cholesterol, hypertension, and glucose were all treated in an intensive and structured way. The conservatively treated patients in the Steno-2 trial had a 50% death rate over 13 years—a shocking mortality rate.

The big difference in the glucose management compared with ACCORD was that Steno-2 provided an evidence-based protocol consistent with best practice. Overweight diabetics in Steno 2 were given Metformin. Gliclazide was added to the medical therapy of obese patients who did not achieve goal glucose reduction. If the glicazide-Metformin combination did not result in reduction of A1c to 7.0, then gliclazide was dropped and long-acting insulin added to the Metformin. This is a critical difference.

The insulin-Metformin regimen does not cause weight gain. Metformin is the only diabetic drug with powerful evidence of reduction of incidence of myocardial infarction (39%) and all-cause mortality (42%) compared with diet and exercise. Pioglitazone has some evidence of a less potent effect in this regard. If Metformin fails, it is because the patient is not making enough insulin and so, simply replacing the insulin that the patient cannot make should result in preservation of the protective effects of Metformin regarding heart attack and death. Steno 2 seems to bear that out.

Metformin is a drug that has powerful vascular effects on a par with a statin or ACE inhibitor. Metformin is associated with modest weight loss, decreased total cholesterol, decreased triglycerides, decreased LDL cholesterol, improved endothelial function (increased ability of the artery to dilate), and positive effects on the increased tendency to clot formation found in diabetics. Metformin cuts to the center of the metabolic syndrome and diabetes by upregulating AMPK-its only site of action. A study abstract that has just been published by David Lefer’s group shows that a single dose of metformin reduces myocardial infarction size in lab animals by 50% by upregulating AMPK and nitric oxide activity.

A single dose of statin does the same thing. These are powerful effects. The National Registry of Myocardial Infarction showed that the patient entering an emergency room with a heart attack was one-third as likely to develop CHF or die if a statin was started in the ER and continued during the hospitalization. These effects are due most likely to a reduction in myocardial infarction size.

These two studies, ACCORD and Steno 2 teach us a great deal about the way that we should manage the epidemic of type 2 diabetes. It is not about the sugar. Simply driving the sugar lower may well be harmful. The Institute of Medicine in 2001 recommended evidence-based protocols consistent with best practices in the treatment of chronic conditions. Steno 2 fits that recommendation and ACCORD clearly does not. Steno 2 should provide us with the model of the future—that is a coordinated integrated attack on global cardiovascular risk using medications that have been shown to reduce the risk of heart attack and death. ACCORD has shown us that aggressive care using any modality that lowers the risk factor may be dangerous and a premium should be placed on medications that have been shown to reduce events and death. The ADA treatment guidelines say that Metformin should be started at diagnosis in the absence of contraindication and these results underscore that recommendation. Long-acting insulin should be used sooner as in Steno 2 rather than later.

Over a 10-year period in South Carolina ending in 2007, I personally treated 450 type 2 diabetic patients with an average age of 65 and multiple comorbidities. By using Metformin, a self-adjusted long-acting insulin shot and the South Beach diet, I was able to maintain an A1c in that patient population that ranged from 6.8-7.0 over the entire 10-year period. These patients did very well with nothing approaching a 50% mortality. Most of these patients had hypertension and cholesterol problems. They were treated in a systematic protocol-driven manner. Pharmaceutical interventions were chosen using the following cascade of priorities.

  1. Event reduction
  2. Effect on weight
  3. Beneficial effects on other risk factors
  4. Effects on endothelial function and arterial structure
  5. Convenience
  6. Side effect profile
  7. Reduction of insulin resistance

The basic protocol used looked like this:

ACE/ARB Statin Metformin
Amlodipine long-acting insulin

Most of these drugs are now available at Walmart for $4 a month. Most of the patients actually lost weight and kept it off. The patients could be seen at a reasonable rate of speed and for relatively little cost. Aggressive diabetes management can be an important part of vascular risk reduction.

The last message anyone should take away from the ACCORD shut-down is that aggressive treatment is a bad idea. Evidence of event reduction is king and other treatments should only be used when the patient cannot take the evidence-based therapy.