On Monday, the National Heart, Lung, and Blood Institute at the National Institutes of Health shut down part of a $300 million study called ACCORD that had been following more than 10,000 very high-risk diabetic patients at 77 medical centers for as long as seven years. All had not only long-standing diabetes but also two or more other major cardiovascular risks such as obesity, smoking, or a previous heart-disease history. An unexpectedly large number of deaths had occurred among patients who were being aggressively treated to bring their blood sugar level below currently recommended standards. There were 257 deaths, about 25 percent higher than the 203 deaths in patients who got the usual care. (Because all patients were carefully monitored and managed, the annual death rate for both groups, about 1.4 percent and 1.1 percent respectively, was far below the 5 percent typical of other patients at similarly high risk.)
Diabetes is hard on the heart and the blood vessels. The rates of strokes and heart attacks and of microvascular complications like blindness, kidney failure, and nerve damage are far higher than in those free of the disease. To lower the risk, doctors constantly urge patients to push their blood sugar down and keep it under control. The desirable target for most is thought to be a blood glucose level (termed A1C level) of below 7 percent, compared with about 4 to 6 percent in the nondiabetic population. An important question that one part of the ACCORD trial was trying to answer was whether achieving a goal of below 6 percent could bring down the number of major heart events. So half of the patients were assigned a target A1C of below 6 percent and got an intensive glucose-lowering regimen of medication and close management; the other half got standard care and a typical target of 7 to 7.9 percent.
Based on the disturbing findings, the NHLBI yesterday recommended that patients like those in the trial not strive for an A1C level below 7 percent. Martin Abrahamson, medical director of the Joslin Diabetes Center in Boston and associate professor of medicine at Harvard Medical School, agrees with that advice. "An A1C of below 6 percent is not and has never been a treatment guideline, and we should not advocate it now," he says.
What is most important, says Abrahamson, is to push and keep pushing to get blood glucose levels down at least to some extent. "One point is quite a lot," he says. Indeed, studies have shown that every 1-point drop in A1C reduces the risk of eye, kidney, and nerve damage by 40 percent. The ACCORD study was not designed to take such complications into account—only to evaluate the effect on heart events and deaths.
The experts involved with the ACCORD trial said yesterday that they don't have an explanation for what happened. The physicians at the 77 centers were free to use the drugs of their choice. Many prescribed Avandia, and because of last year's concern over cardiovascular risks that the medication may pose, researchers already have looked for a possible link. None has been found. NHLBI Director Elizabeth Nabel promised that the data will soon be published in a peer-reviewed journal, but only as a report—not as an explanation.
— Avery Comarow