New ACCORD Insights: Steady Glycemic Control Most Important

The Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial enrolled 10,251 participants who had Type 2 diabetes and either a previous cardiovascular event or other evidence of high cardiovascular risk. Patients were randomly assigned to either an intensive glycemic strategy with the aim of achieving hemoglobin A1c levels below 6.0% or a standard strategy with the aim of keeping A1c between 7.0% and 7.9%.

Participants had A1c measurements taken every 4 months. The primary end point of the ACCORD trial was a composite of cardiovascular mortality, nonfatal myocardial infarction, or nonfatal stroke. All-cause mortality was a predefined secondary end point.

The intensive treatment strategy was stopped early when it was determined that all-cause mortality was higher in that group. The current post hoc analysis was conducted to shed light on this unexpected finding. The investigators tested patients' A1c levels over the 3.4 years of follow-up, and then compared 4 measures to determine whether any predicted all-cause mortality: mean A1c levels over the follow-up period, final A1c, decrease in A1c in the first year, and decrease in A1c in the first 4 months.

Both before and after adjustment for other covariates, a higher average A1c level was a stronger predictor of mortality than either the final A1c or the decreases in A1c in the first year. Higher average A1c level was associated with higher risk for death: Each percentage-point increase in mean A1c was associated with an approximate 20% increase in all-cause mortality. The risk for death with the intensive strategy increased almost linearly from A1c values of 6.0%-9.0%. The original ACCORD finding that mortality risk was higher with the intensive strategy than with the standard strategy appeared to hold only when average A1c level was above 7%.

When ACCORD halted its glycemic control study after indications that intensive control might be harmful and when the Action in Diabetes and Vascular Disease (ADVANCE) study and the Veteran's Affairs Diabetes Trial (VADT) also failed to show benefit of tight glycemic control, concern arose that clinicians might begin to discount A1c targets. That concern was sufficient to prompt a position statement from the American Diabetes Association, the American College of Cardiology Foundation, and the American Heart Association, reaffirming the general A1c goal of < 7%.

For many, the negative findings from the "big 3" trials did not fit our understanding of diabetes, glycemic control, or risk for complications and death, yet as scientists, we must consider the evidence. Thus, despite the limitations of post hoc analysis, the finding that risk for death increased more or less linearly with increasing A1c is reassuring as well as consistent with other studies.

It is important to note that average A1c level was a better predictor of mortality than decline in A1c, because one of the concerns with the ACCORD trial was that the rapid reduction in A1c may have been responsible for the excess mortality. On the contrary, the current analysis suggests that failure to respond (rather than a good response) to intensive therapy might somehow be associated with higher risk for death. This is a potentially important hypothesis that needs further exploration.

Another topic for further study involves long-term glycemic control. Patients with diabetes need ongoing therapy adjustments to maintain glycemic control, yet those adjustments often do not occur in a timely manner. As a result, patients may experience long periods of inadequately controlled A1c. It may be that good glycemic control over the duration of diabetes as opposed to tight control interspersed with spikes in A1c is the key to risk reduction. Some hint of this was seen in the legacy effect observed in the long-term follow-up of the United Kingdom Prospective Diabetes Study (UKPDS). Slow and steady might well win the race.

Diabetes Care 2010; 33(5):983-90