2011 HSR&D National Meeting Abstract
3017 — Hemoglobin A1c as a Predictor of Incident Diabetes
Conlin PR (VA Boston Healthcare System), Cheng P
(Tampa VAMC), Neugaard B
(Tampa VAMC), Foulis P
Several studies have suggested HbA1c (A1c) levels may predict incident diabetes. With new recommendations for use of A1c in diagnosing diabetes, many patients with A1c results below the diagnostic threshold will be identified. Clinicians will need to categorize risk for a subsequent diabetes diagnosis in such patients. The objective of this study was to determine the ability of A1c to predict a subsequent diagnosis of diabetes.
The study employed a retrospective chart review using electronic medical record data from two Department of Veterans Affairs Medical Centers. Patients were identified with no prior diagnosis of diabetes but having a baseline A1c test between January 2000 and December 2001. These patients were tracked for 8 years for a subsequent diagnosis of diabetes.
12,375 patients were identified with a baseline A1c < 6.5% and without a diagnosis of diabetes. During an average follow-up period of 4.4 years, 3,329 (26.9%) developed diabetes. A1c > = 5.0% carried a significant risk for developing diabetes during follow-up. When compared to the reference group (A1c < 4.5%), A1c increments of 0.5% between 5.0 – 6.5% had adjusted odds ratios of 1.70 (5.0-5.4%), 4.87 (5-5.9%), and 16.06 (6.0-6.4%), (p < 0.0001). Estimates of hazard ratios similarly showed significant increases for A1c > = 5.0%. A risk model for incident diabetes within 5 years was developed and validated employing A1c, age, body mass index, and systolic blood pressure.
The incidence of diabetes is progressively and significantly increased among patients with an A1c > = 5.0%, with substantially expanded risk for those with A1c 6.0-6.4%.
For individuals without a diagnosis of diabetes, A1c along with other clinical factors, can be used to predict the 5-year incidence of diabetes