3104 — Characterizing Veterans with Diabetes who are Missed by A1c Screening
Jackson SL, Emory University, Division of Biological and Biomedical Sciences; Olson DE, Atlanta VA and Emory University School of Medicine; Mohan A, Emory University School of Medicine; Tomolo A, Atlanta VA and Emory University School of Medicine; Barb D, Emory University School of Medicine; Zhu M, and Long Q, Emory University, Rollins School of Public Health; Phillips LS, Atlanta VA and Emory University School of Medicine;
Although screening to detect unrecognized dysglycemia is recommended, best implementation of screening with available tests to translate current knowledge on hemoglobin A1c (A1c) is unknown. Since the quality of primary care in the VA is high, we examined opportunistic screening with A1c during VA primary care visits, and identified factors associated with being “missed” by A1c to inform VA policy on screening for diabetes.
Screening was offered to patients meeting NIDDK/ADA guidelines: without known diabetes, and with age >=45 yr or BMI >=25 kg/m2 with another risk factor. An OGTT identified hyperglycemia classified by ADA criteria, and A1c findings were evaluated according to ADA (diabetes if A1c >=6.5%), and the part of the VA/DoD (diabetes if A1c >=7.0%) guidelines limited to A1c alone.
Of 790 screened Veterans who completed OGTT and A1c, 9.9% had diabetes. 91% were male, 71% black, and 29% white, with average age 60 years and BMI 32. Screening with A1c at the ADA cutoff detected 27% of patients with diabetes found by OGTT, while screening at the VA cutoff detected only 12%. Of the 78 with diabetes by OGTT, A1c at the ADA cutoff was more likely to miss whites (OR 2.37, p = 0.051), but less likely to miss subjects with higher BMI (OR 0.86, p = .042), greater age (OR 0.92, p = .062), greater fasting glucose (OR 0.98, p = 0.048), and greater 2-hour glucose (OR 0.98, p = 0.01). Results for BMI, fasting glucose, and 2-hour glucose remained significant in analyses using the VA cutoff. No significant differences in blood pressure or lipids were observed between patients with diabetes who were missed vs. detected by A1c, despite fasting and 2-hour glucose levels being significantly lower.
Many Veterans targeted by A1c testing still have unrecognized diabetes, and Alc screening may systematically miss certain patients who would benefit from more stringent control of other comorbidities.
Efforts should be made to identify patients with diabetes who are likely to be missed by A1c screening. This would enable VA providers to implement care for diabetes and associated co-morbidities to prevent future complications, but would require more rigorous screening than with A1c alone.