BOSTON — It's still not clear why mortality was higher with the intensive glycemic control strategy of the ACCORD trial, but new analyses of the data highlight the fact that higher glucose levels, not lower, remain the most likely culprit.
At a special evening session held during the annual meeting of the American Association of Clinical Endocrinologists, researchers from the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial summarized results of recent subanalyses of the trial data and suggested ways in which the new findings, while still lacking an ultimate conclusion, might nonetheless help inform clinical decisions—at least among patients who resemble the study population.
In ACCORD, all-cause mortality was greater among patients randomized to intensive glycemic control, with the aim of getting patients to a hemoglobin A1c below 6%. When the glycemic arm of the study was stopped early, at 3.4 years rather than the planned 5.6 years of follow-up, the hazard ratio was 1.22, compared with standard treatment (N. Engl. J. Med. 2008;358:2545–59).
Panel moderator Dr. Faramarz Ismail-Beigi said that the overall conclusion from the glycemia arm of the ACCORD trial pertains only to that patient group: In older patients with a longer duration of diabetes and established cardiovascular risk factors, attempting to achieve normoglycemia does not reduce all-cause or cardiovascular mortality and may increase the risk.
However, he said, it's important to remember that the 10,194 patients enrolled in ACCORD did not represent the entire type 2 diabetes population. The ACCORD patients had an average age of 62 years, a 10-year duration of diabetes, and a median HbA1c of 8.1%. A third had experienced prior cardiovascular events.
“This cohort represents a little bit less than half of all U.S. patients with type 2 diabetes. So it represents a large group of people, but not everybody. We're not talking about people who are newly diagnosed with diabetes, middle-aged, or younger,” said Dr. Ismail-Beigi, professor of medicine at Case Western Reserve University, Cleveland.
Several possible mechanisms to explain the results were put forward at the time the glycemia arm of ACCORD was stopped in 2008, including hypoglycemia, weight gain, individual drugs, drug combinations, or the rapid reduction of glucose levels early in the trial. But now, new data refute some of these hypotheses.
Dr. Elizabeth R. Seaquist, professor of medicine and director of the Center for Diabetes Research at the University of Minnesota, Minneapolis, presented just-published data showing that the excess risk of all-cause mortality associated with intensive treatment in ACCORD was associated with persistently high HbA1c rather than low HbA1c, regardless of treatment group assignment.
Average HbA1c was the strongest predictor of death for both groups: A 1-percentage-point increase was linked with a 22% increase in mortality, after adjustment for a variety of potentially confounding baseline factors. When each group was examined separately, the relationship between HbA1c and death was much stronger among those in the intensive treatment group, with a statistically significant 66% increase in all-cause mortality for every 1-percentage-point higher HbA1c vs. a nonsignificant 14% increase for the standard treatment group.
The greatest excess risk of death associated with the intensive treatment group occurred among the patients whose average HbA1c remained above 7% despite their treatment assignment. In the intensive treatment group, there was a steady increase in mortality as the HbA1c rose from 6% to 9%, whereas no such relationship was seen in the standard treatment group. The excess mortality in the intensive group was seen only at an HbA1c above 7%, not below, Dr. Seaquist reported.
The relationship between mortality and the last HbA1c recorded before death and the decrease in HbA1c over the first year did not differ between the two groups, suggesting that the rate of change in HbA1c from baseline was not associated with increased risk of death. “These analyses do not support the view that rapid reduction of glucose levels or lower average A1c independent of other factors led to the excess risk of death,” she said.
Dr. Saul Genuth, professor of medicine at Case Western Reserve University, summarized findings from two studies published last year suggesting that severe hypoglycemia was a risk factor for increased mortality in ACCORD, but that the relationship between hypoglycemia and mortality did not explain the difference in outcomes between the intensive and standard treatment groups.
Patients with poorer glycemic control had a greater risk for hypoglycemia in both groups, and among those who experienced severe hypoglycemia, the risk of death was actually lower in the intensive treatment arm, he said.
The frequency of severe hypoglycemia events requiring medical assistance was 4.3 per 100 person-years for the intensive arm, compared with 1.4 for standard treatment. There was a slow decline in severe hypoglycemic events over the 3.4 years of the trial in the intensive group, whereas the rate remained steady in the standard treatment group. “This should encourage all of us to keep educating our patients about preventing hypoglycemia,” Dr. Genuth said.