Hemoglobin A1c: It’s Not Just for Diabetics Anymore

Abstract & Commentary

Synopsis: The relative risk for each 1% increase in hemoglobin A1c was 1.3 for CHD, 1.2 for CVD, and 1.2 for all cause mortality.

Source: Khaw KT, et al. Ann Int Med. 2004;141:413-420.

This prospective cohort trial assessed the risk of cardiovascular disease in 25,623 patients with 6 levels of a hemoglobin A1c level obtained at baseline: < 5% (n = 2766), 5-5.4% (n = 3573), 5.5-5.9% (n = 2531), 6-6.4% (n = 813), 6.5 to 6.9%; n = 157), > 7% (n = 149). One percent (n = 243) of the study had been previously diagnosed with diabetes (mean HgbA1c = 8%). Participants were followed over 6 years for 3 end points: coronary heart disease events, cardiovascular disease events and all-cause mortality.

In men, all end points increased in frequency as the hemoglobin A1c increased, with the most significant jump occurring at the 6.5-6.9% category (RR = 3.4 CHD; 3.0 CVD; 3.5 all cause mortality). There was a progressive increase in events in women beginning at the 6-6.4% category (RR = 2.3 CHD; 1.6 CVD; 1.6 all cause mortality). There was a substantial increase in all events in both sexes at the > 7% category (men, RR = 7 CHD; 5 CVD; 3.4 all cause mortality; women, RR =4.7 CHD; 8.0 CVD; 6.9 all cause mortality).

After adjustment for age and risk factors, the relative risk for each 1% increase in hemoglobin A1c was 1.3 for CHD, 1.2 for CVD, and 1.2 for all cause mortality. In the known diabetic subgroup, the risk for CHD, CVD and mortality (RR = 2.4, 2.2, and 2.3) was reduced to 1.0 after adjustment for hemoglobin A1c.

Comment by Jeff Wiese, MD

This study confirms previous trials that found a continuous risk of macrovascular disease as the hemoglobin A1c increases.1,2 More importantly, this study suggests that the risk for cardiovascular events in diabetics is mediated entirely by the hemoglobin A1c. In the diabetic cohort in the study, the risk for all events was reduced to that of non-diabetics once the model was adjusted for hemoglobin 1Ac levels.

Hemoglobin A1c is a marker for systemic glycosolation due to elevated blood glucose levels over the previous 3 months. Previous trials have suggested that tight glucose control decreases the risk for complications from diabetes; this and other trials suggest that tight glucose control may also reduce the risk of cardiovascular events in non-diabetic patients (by 20 to 30% for each percent decline in the HgbA1c).3

The results of this study should be interpreted with caution; however, as the cohort design permits commentary on association of hemoglobin A1c with cardiovascular events, but not causation. The latter claim will require a randomized trial in which non-diabetic patients are randomized to tight glycemic control vs status quo. There are numerous confounders that may explain the increased risk of cardiovascular events as the hemoglobin A1c increases. For example, the percentage of patients who smoked cigarettes also progressively increased with each hemoglobin A1c tier, as did the body mass index, suggesting that obesity, tobacco or inactivity may have been responsible for the increased mortality. Although Khaw and colleagues adjusted for these risk factors, the tight relationship between obesity and hemoglobin A1c makes complete adjustment impossible. Physicians tempted to tightly control hemoglobin A1c levels in non-diabetics, while neglecting the obesity that caused the increase in hemoglobin A1c, may find the results to be less than desired. Nonetheless, should a causal link between hemoglobin A1c and cardiovascular mortality be established in non-diabetic patients, the implications could be huge. Khaw et al estimate that a society-wide reduction of hemoglobin A1c of 0.1% has the potential to reduce the total US mortality by up to 6 percent.

Dr. Wiese, Chief of Medicine, Charity, and University Hospitals, Associate Chairman of Medicine, Tulane Health Sciences Center, is Associate Editor of Internal Medicine Alert.


1. Coutinho M, et al. Diabetes Care. 1999;22:233-240.

2. Gerstein HC. Diabetic Med. 1997;14 (Suppl. 3):S25-S31.

3. Park S, et al. Diabetes Care. 1996;19:450-456.