HbA1c Predicts Cardiovascular Risk

HbA1c had a greater impact on predicting cardiovascular disease (CVD) in women (P<0.001), but prediction also improved significantly in men (P=0.039). 

The results could prove useful to ongoing discussions about managing CVD risk in patients with diabetes, investigators reported online in Archives of Internal Medicine. 

Nina P. Paynter, PhD, of Brigham and Women's Hospital in Boston, and co-authors wrote in a discussion of their findings, "These results may be particularly helpful in light of current discussion around treatment choices for diabetic patients for prevention of CVD, including use of statins and aspirin."  "Our results suggest that the use of HbA1c levels as part of overall CVD risk scores may improve predictive ability in diabetic patients, whose HbA1c levels are routinely measured in clinical practice." 

Diabetes has a well-established association with CVD risk. Whether the association varies by HbA1c levels or other CVD risk factors has not been determined. The issue has clinical relevance, the authors wrote, because if HbA1c improved CVD risk prediction, a risk model that included an HbA1c term for diabetes could be used both for patients who have the disease and those who don't. 

To examine the question, the authors analyzed data on 24,674 women (685 of whom had diabetes) who participated in the Women's Health Study (WHS) and 11,280 men (563 with diabetes) from the Physicians' Health Study (PHS). The WHS participants had a median follow-up of 10.2 years, and the PHS participants had a median follow-up of 11.8 years. 

During follow-up in the WHS, 125 CVD events occurred in the women with diabetes and 666 in those who did not have the disease. In the PHS the 563 men with diabetes had 170 cardiovascular events versus 1,382 in the nondiabetic men. Diabetes was considered a CVD risk equivalent.

Paynter and colleagues analyzed the effect of adding HbA1c to the 10-year risk estimates of the National Cholesterol Education Program for Adults (ATP III) and the Reynolds Risk Score (RRS), which includes C-reactive protein and parental history of premature myocardial infarction.

The ATP III guidelines define diabetes as a CVD equivalent, meaning that all patients with diabetes fall into the highest-risk category.

Overall, models that included HbA1c significantly improved discriminatory power (C statistic 0.177, P<0.001) and led to net reclassification improvement (NRI) of 26.7% for the ATP III (P=0.001) and 23.6% for RRS (P=0.003) among diabetic participants in the WHS. 

Including HbA1c levels had a more modest impact among diabetic men in the PHS but still yielded a significant improvement in discriminatory power in the ATP III model (C statistic 0.039, P=0.02) and NRIs of 9.2% for ATP III (P=0.04) and 12.4% for RRS (P=0.004).  

HbA1c levels also improved prediction as compared with a dichotomous term for diabetes (versus a risk equivalent) in women (NRI 11.8%, P=0.03) but not in men. 

The study had several limitations that were highlighted in an accompanying editorial. Although the analysis involved almost 36,000 men and women, the number of diabetic participants (1,248) was modest, and most of them had well-controlled diabetes, as reflected in an HbA1c of 6.5% to 7.0%, noted Mark J. Pletcher, MD, of the University of California San Francisco.

Whether enough diabetic participants with elevated HbA1c levels were included to ensure accurate modeling is unclear, he continued. Additionally, he noted, the analysis did not show that including HbA1c in a risk question is superior to use of an indicator variable for diabetes (such as 1 or -1).

Despite the limitations, the analysis demonstrated the value of allowing reclassification of diabetic patients by means of HbA1c, as opposed to the practice of including all diabetic patients in the highest-risk category for CVD, Pletcher wrote.

• Explain that the addition of HbA1c to models improved prediction of risk for cardiovascular events compared with using diabetes to put all patients in the highest risk category.

• Note that the actual number of diabetic patients in the large cohorts was small and most had well-controlled diabetes.

Paynter NP, et al "Cardiovascular risk prediction in diabetic men and women using hemoglobin A1c versus diabetes as a high-risk equivalent" Arch Intern Med 2011; DOI:10.1001/archinternmed.2011.351.