Comment on Hempe et al. The Hemoglobin Glycation Index Identifies Subpopulations With Harms or Benefits From Intensive Treatment in the ACCORD Trial. Diabetes Care 2015;38:1067–1074

Hempe et al. (1) analyzed the Action to Control Cardiovascular Risk in Diabetes (ACCORD) population to assess whether a measure based on the relationship between glycated hemoglobin (HbA_1c) and fasting plasma glucose (FPG) at baseline (termed a “hemoglobin glycation index”) was associated with cardiovascular outcomes, death, and/or hypoglycemia. They found that intensive treatment reduced the risk of the primary composite cardiovascular end point in people whose index fell in the middle and lower third (representing a lower HbA_1c relative to FPG) and had a neutral effect in people with the highest index. They also found that the reported excess risk of death with intensive treatment might have been confined to people with the highest index and concluded that “HbA_1c is not a one-size-fits-all indicator” of glucose control (1).

We congratulate the authors for this interesting analysis based on their prior studies of rapid glycation as a possible contributor to the variation in the relationship between HbA_1c levels and adverse outcomes. Although limited (like all post hoc subgroup analyses) to generating hypotheses rather than providing firm answers, these findings are consistent with prior analyses suggesting that individuals in ACCORD were at increased risk during intensive glycemic treatment if they 1) previously had HbA_1c >8.5% during ongoing treatment (2), 2) failed to reduce HbA_1c more than 0.5% soon after intensification of treatment (3), or 3) continued to have HbA_1c >7.0% thereafter (3)—observations that all signify an inability to control HbA_1c despite usual treatment tactics. A common explanation for an HbA_1c higher than predicted by the concurrent FPG is imperfect control of daytime glucose levels. Indeed, use of oral glucose-lowering agents or insulin can alter the relation of HbA_1c to FPG. Thus, an analysis from >12,000 participants in the Outcome Reduction With Initial Glargine Intervention (ORIGIN) trial showed that when the FPG was ∼175 mg/dL (as at baseline in ACCORD), HbA_1c levels were ∼0.2% higher in the presence of oral agents, compared with no treatment (4). In a different report comprising >1,200 people starting basal insulin, HbA_1c declined from 8.71 to 7.02% (∼20% decrease) in response to a fall in FPG from ∼195 to 120 mg/dL (∼40% decrease) and there was little change in postprandial excursions (5). These reports illustrate that glycemic therapy unrelated to inherent glycation rates may also lead to an HbA_1c higher than predicted by FPG. As ACCORD participants with higher indices had a longer duration of diabetes and often used insulin (1), they likely included many individuals with poor postprandial glycemic control despite multiple therapies. Outcomes in this subgroup may therefore be related to physiologic abnormalities underlying postprandial hyperglycemia, side effects of treatments seeking to reduce postprandial increments, or glycemic variability itself.

On the basis of these considerations, we suggest that “hemoglobin glycation index” may not be an ideal name for this measure of risk, as it suggests a nonmodifiable abnormality of glycation. Instead, a term such as “A_1c-FPG index” would be more comprehensive, including the possibility of a contribution from daytime hyperglycemia, which is potentially modifiable, in addition to rapid glycation. The index could be evaluated as a clinically helpful trigger for reassessment of both glycemic targets and treatment tactics for individual patients.