The U.S. Preventive Services Task Force (USPSTF) recommends diabetes screening every three years for patients with hypertension because they have an increased risk of developing diabetes. Experts do not agree on the usefulness of screening because no study has found early intervention to be more effective than intervention when symptoms appear. The role of hemoglobin A1C screening is unclear, but some evidence suggests that the use of A1C levels is more sensitive than fasting blood glucose levels. In addition, A1C levels can define treatment goals and risk of complications in patients with established disease. Edelman and colleagues examined whether A1C levels could predict new-onset diabetes in an outpatient population.

Patients 45 to 64 years of age without hypertension or diabetes were enrolled in the study. A1C levels were measured at baseline, and patients were interviewed annually for two years and were rescreened at year 3 to determine whether they had developed diabetes (defined by self-report, an A1C level of 7 percent or higher, or a fasting blood glucose level of 126 mg per dL [7 mmol per L] or higher).

Of the 957 patients who completed the follow-up screening, 73 developed diabetes. None of these patients had a baseline A1C level of 4.5 percent or lower. Diabetes incidence steadily increased with increasing baseline A1C level. The annual incidence rate was 0.8 percent for patients with normal A1C levels (5.5 percent or lower), 2.5 percent for patients with high-normal levels (5.6 to 6.0 percent), and 7.8 percent for patients with elevated levels (6.1 to 6.9 percent). After considering other possible associated characteristics, the authors found that only baseline body mass index was additionally associated with increased diabetes risk, with obese patients at the highest risk.

The authors conclude that A1C levels can identify patients at high risk for diabetes. Obese patients are particularly at risk. The authors suggest that physicians screen their patients, paying special attention to those with risk factors such as hypertension, obesity, and a family history of diabetes. Identifying patients with A1C levels from 6.1 to 6.9 percent might be beneficial in terms of lifestyle counseling or pharmacotherapy to delay onset of diabetes. This study did not address whether A1C levels could identify more at-risk patients and offer better prevention than identification based on other risk factors such as obesity. The effectiveness and cost-effectiveness of screening also are not known.

editors note: The U.S. Preventive Services Task Force¹ gives a B recommendation to screening high-risk patients (i.e., those with hypertension or hyperlipidemia) for diabetes but states that there is insufficient evidence to recommend universal screening. No randomized controlled trials have compared outcomes based on diabetes screening with outcomes based on symptom detection. Contrary to this article, hemoglobin A1C testing is described as relatively insensitive to lower-level elevations of fasting blood glucose: 87 percent of patients with glucose intolerance determined by fasting blood glucose levels also have a normal A1C level (i.e., less than 6.1 percent).—c.w.

1U.S. Preventive Services Task Force. Guide to clinical preventive services. 3d ed. Washington, D.C.: Agency for Healthcare Research and Quality, 2003.