Refining HbA1c Targets
ACCORD Follow-Up Reinforces Caution
By Bill Malone
The Action to Control Cardiovascular Disease Risk in Diabetes (ACCORD) trial compared intensive versus standard glycemic control in older patients with type 2 diabetes and cardiovascular disease risk factors. The study was abruptly stopped in 2007 when mortality was significantly higher in the intensive treatment group than in the standard treatment group (5% vs. 4%). This issue of Strategies explores a recent paper from the ACCORD study group that reports on outcomes after 5 years, 1 ½ years after intensive treatment was relaxed.
American Diabetes Association (ADA) guidelines recommend a glycated hemoglobin (HbA1c) target of <7% for type 2 diabetics. However, many in the medical community have questioned whether more aggressive targets that would push down HbA1c to into the normal range might not help improve cardiovascular disease (CVD) risk. The landmark ACCORD trial, which included more than 10,000 older patients with long-standing type 2 diabetes and CVD risk factors, attempted to answer this question by randomizing participants to either of two treatment strategies, one that targeted normal HbA1c <6%, and the other targeting a higher range, 7–7.9%. Now, even after the intensive treatment arm was halted due to excess mortality, the ACCORD study group has published a paper showing that this unexpected trend continued, with a reduction in nonfatal myocardial infarction (MI) in the intensive treatment group, but overall increased mortality of 7.6% versus 6.4% (N Engl J Med 2011; 364:818–828).
Before the HbA1c target for the intensive treatment arm was relaxed, this group was similar to the standard treatment group in the rate of the primary outcome, a composite of non-fatal MI, nonfatal stroke, or death from cardiovascular causes. However, the intensive treatment group had more deaths from any cause—primarily cardiovascular—and fewer nonfatal MIs. During the full follow-up period, this trend persisted. From randomization until the end of the study, 391 intensive treatment subjects died, compared to 327 deaths in the group that spent the entire time under standard treatment. After being switched to standard treatment, the median HbA1c in the intensive treatment group rose from 6.4% to 7.2%.
Most puzzling is that in the study, two of the key components of the primary outcome essentially go in opposite directions, with cardiovascular death going one direction and nonfatal MI going in the other, a finding yet to be explained by the new study’s further analysis of the data, noted lead author Hertzel Gerstein, MD. “We don’t yet know the reason for this, and we may never know the reason, because the study was not designed to answer those questions,” he said. “What it does tell us is that because of safety concerns, we can’t justify recommending that older people with CVD risks and long-standing type 2 diabetes be treated intensively to target an HbA1c that’s in the non-diabetic range, less than six percent. That was really what ACCORD tested.” Gerstein is professor of medicine and director of the division of endocrinology and metabolism at McMaster University in Hamilton, Ontario.
Since the intensive treatment arm was halted and preliminary results published in 2008, controversy over ACCORD has centered on speculation about why the intensive treatment arm experienced higher mortality—especially since on other measures these people fared better, including fatal and nonfatal MI, fatal or nonfatal stroke, unstable angina, revascularization, and hospitalization for congestive heart failure. Different drug combinations, weight gain, hypoglycemia, and other theories have been advanced to explain the outcome.
However, sorting out all of these variables to look for a cause might never be possible, because determination of the particular hypoglycemic drugs that were prescribed or how often a person checked his or her blood glucose at home occurred after randomization and, as such, confounds accurate analysis, Gerstein said. “When trying to understand the impact of a drug that you gave after a randomization—which you may have given because of how they responded to what they were already taking—it’s hard to disentangle the actual effect of the drug from the reason that you gave the drug,” he explained. “It’s very difficult to tease out the relationships. All we can do with these post-randomization analyses is explore and create hypotheses. Even something really clear, for example if people with a certain drug had experienced a huge benefit, that would still just be a hypothesis, and you’d need a prospective trial.”
To resolve the controversy over the benefit of more intensive glycemic control, further studies will be needed, contended James Boyd, MD, associate professor of pathology at the University of Virginia School of Medicine, who was not involved in the study. “Certainly the benefits in the intensive treatment group in lowering cardiovascular risks and microvascular complications would be evidence that you don’t want to just throw aside and automatically opt instead for a higher goal glucose concentration,” he said. “There are so many complexities and so many residual questions, that it’s hard to say going with a higher glucose goal can just be accepted without further study.”
In addition to the varieties of multiple drug treatment regimens used to target lower HbA1c in the study, Boyd also noted that with HbA1c as the target, fluctuations in glucose are unknown and unaccounted for in the study. “HbA1c remains a good, integrated marker of long-term glycemic control, but it doesn’t tell you anything about fluctuations of glucose that may have occurred over the four to six week period that HbA1c averages,” he said.
New, more detailed analyses from the ACCORD study group are forthcoming. Meanwhile, the study should not be interpreted to counter current ADA HbA1c guidelines for a <7% target, according Gerstein. “Actually, ACCORD tells us nothing about the safety of targeting less than seven percent. That’s an important principle. So we should stick with the ADA guidelines and not try to get super control, just good control,” he said. “The most important thing is that the ACCORD study adds a tremendous fund of knowledge to our management of patients. And it raises as many questions as it answers, as any good trial can do. But we should not interpret ACCORD as an excuse for therapeutic nihilism in the management of diabetes. That’s a clear danger, because ACCORD did not randomize people to good versus poor control; it randomized people to perfect glycemia versus good control. That’s an important distinction to make.”
Join us for a discussion of this article, and interact with your peers, on LinkedIn.