Elevated HbA1c Associated With Nonfatal CVD, Even in Nondiabetics

Shelley Wood

September 18, 2007 (Amsterdam, the Netherlands) - Even in nondiabetics, increased HbA1c levels are associated with a significantly increased risk of nonfatal cardiovascular disease after other cardiovascular risk factors are accounted for, an analysis from the Hoorn cohort study shows.
While HbA1c is a well-established predictor of cardiovascular events in diabetics, this is the first study to link elevated HbA1c to fatal and nonfatal cardiovascular disease in nondiabetics, Dr Esther van 't Riet (VU University Medical Centre, Amsterdam, the Netherlands) reported here at the European Association for the Study of Diabetes 2007 Meeting.
"The clinical meaning of this is that, even in subjects without diabetes, it is very important to maintain optimal glycemic control," van 't Riet told heartwire. "Trying to lower HbA1c may not be the first thing you do in terms of treatment, but you can see that when you have subjects without diabetes who have high HbA1c levels there is some degree of risk, and when you lower HBA1c you lower their risk of CVD. This is evidence of that relationship."
The Hoorn study is a population-based cohort analysis that enrolled 2484 subjects back in 1989 to 1990. A total of 1674 nondiabetic subjects with data on baseline HbA1c were included in the current analysis and analyzed by tertiles of baseline HbA1c. As van 't Riet reported here, subjects in the highest HbA1c tertile (>5.6%) had significantly increased risk of developing nonfatal cardiovascular disease or dying of cardiovascular disease, even after adjustment for age and sex: hazard ratios of 2.11 and 1.73, respectively. When the analysis was further adjusted for cardiovascular risk factors (hypertension, smoking, LDL, triglycerides, and waist-to-hip ratio), a high HbA1c was still significantly associated with nonfatal CVD (but not fatal CVD), with a hazard ratio of 1.71.
Fasting glucose levels at baseline and measures of two-hour plasma glucose were not significantly associated with increased CVD risk.
No link to impaired glucose tolerance
To heartwire, van 't Riet said that the lack of an association with fasting glucose levels comes as no surprise: no previous reports have linked fasting glucose to CVD in nondiabetics. More surprising, however, was the lack of association with two-hour plasma glucose levels--a measure of impaired glucose tolerance. "That has been reported before, in several large epidemiological studies," she said, "But I think difference comes from the fact that we excluded all diabetic subjects, whereas earlier reports included diabetic subjects. We did an analysis in which we included diabetic subjects [from the Hoorn cohort] and when we did that, we did find an association with two-hour glucose levels."
Commenting on the study, Dr John S Yudkin (University College London, UK) also observed that he would have expected impaired glucose tolerance to be "a better marker" than HbA1c. He also speculated on the significance of HbA1c as a measurement.
Obviously one big question is whether HbA1c is a mediator of cardiovascular risk or merely a marker. Some of the excess risk seems to disappear when the measured CV risk factors are adjusted for, and those measured markers would also probably couple with other unmeasured ones like microalbuminuria, fibrinogen, and CRP, he proposed.
To heartwire, van 't Riet said that, in her opinion, HbA1c is probably a marker, whereas glucose is more of a mediator. But she reiterated that she didn't think HbA1c would necessarily be a good marker to aggressively target in clinical practice. "When you go to treat an individual to lower their HbA1c to decrease their risk of CVD, you are targeting a hazard ratio of 1.71, but when you treat blood pressure or cholesterol, you have the potential to have a much bigger effect on CVD outcomes. I think it's much better to treat blood pressure and cholesterol than HbA1c in subjects without diabetes."