OBJECTIVE Several studies have suggested that HbA_1c levels may predict incident diabetes. With new recommendations for use of HbA_1c in diagnosing diabetes, many patients with HbA_1c results below the diagnostic threshold will be identified. Clinicians will need to categorize risk for a subsequent diabetic diagnosis in such patients. The objective of this study was to determine the ability of HbA_1c to predict the incidence of a diabetic diagnosis.

RESEARCH DESIGN AND METHODS We performed a historical cohort study using electronic medical record data from two Department of Veterans Affairs Medical Centers. Patients (n = 12,589) were identified with a baseline HbA_1c <6.5% between January 2000 and December 2001 and without a diagnosis of diabetes. Patients (12,375) had at least one subsequent follow-up visit. These patients were tracked for 8 years for a subsequent diagnosis of diabetes.

RESULTS During an average follow-up of 4.4 years, 3,329 (26.9%) developed diabetes. HbA_1c ≥5.0% carried a significant risk for developing diabetes during follow-up. When compared with the reference group (HbA_1c <4.5%), HbA_1c increments of 0.5% between 5.0 and 6.4% had adjusted odds ratios of 1.70 (5.0–5.4%), 4.87 (5.5–5.9%), and 16.06 (6.0–6.4%) (P < 0.0001). Estimates of hazard ratios similarly showed significant increases for HbA_1c ≥5.0%. A risk model for incident diabetes within 5 years was developed and validated using HbA_1c, age, BMI, and systolic blood pressure.

CONCLUSIONS The incidence of diabetes progressively and significantly increased among patients with an HbA_1c ≥5.0%, with substantially expanded risk for those with HbA_1c 6.0–6.4%.