Do Not Diagnose Diabetes Based on a Single A1c of 6.5%
You should repeat the A1c test to confirm the diagnosis before labeling this patient with diabetes, and until confirmed, she remains in the prediabetes category. A single A1c ≥6.5% is not sufficient for diagnosis in the absence of unequivocal hyperglycemia or classic symptoms 1, 2, 3.
Why Confirmation is Required
The American Diabetes Association guidelines are explicit and consistent across multiple years: in the absence of unequivocal hyperglycemia (random glucose ≥200 mg/dL with classic symptoms), any diagnostic test result must be confirmed by repeat testing 1, 2, 3. This is a fundamental principle to rule out laboratory error and account for the inherent variability in all diagnostic tests.
The Confirmation Process
You have several options for confirmation 1, 2, 3:
Preferred approach: Repeat the same test (A1c) because there will be greater likelihood of concurrence. If the repeat A1c is also ≥6.5%, diabetes is confirmed—even if one value is 7.0% and the repeat is 6.8%, the diagnosis is established.
Alternative approach: Use a different test (fasting plasma glucose or 2-hour OGTT). If both the original A1c ≥6.5% AND a second different test meets diabetes criteria (FPG ≥126 mg/dL or 2-hour glucose ≥200 mg/dL), diabetes is confirmed.
If results are discordant: When two different tests give conflicting results, repeat the test that was above the diagnostic threshold. The diagnosis is made based on the confirmed test. For example, if you have A1c ≥6.5% but FPG <126 mg/dL, repeat the A1c. If two A1c results are ≥6.5%, the patient has diabetes regardless of the normal FPG 1, 2, 3.
What If the Repeat Test is Below Threshold?
This is a critical clinical scenario. Due to preanalytic and analytic variability, when you repeat a test that was initially above the diagnostic threshold, the second value may fall below the cut point 1, 2. This is least likely with A1c, somewhat more likely with fasting glucose, and most likely with 2-hour glucose values 1, 2.
Barring laboratory error, such patients likely have test results near the margins of the diagnostic threshold. In this situation, you should follow the patient closely and repeat testing in 3–6 months 1, 2.
Important Caveats About A1c Testing
A1c Has Lower Sensitivity Than Glucose-Based Tests
The A1c cut point of ≥6.5% identifies approximately one-third fewer cases of diabetes compared to fasting glucose ≥126 mg/dL when used for universal screening 1, 4. Research confirms this: studies show A1c ≥6.5% has a sensitivity of only 46.8–60.8% for detecting diabetes defined by OGTT, meaning more than half of diabetic individuals may be missed using A1c alone 5, 6.
When A1c May Be Inaccurate
A1c can be misleading in several conditions 1, 3:
- Hemoglobinopathies (sickle cell trait, thalassemia)
- Conditions with abnormal red cell turnover (pregnancy, recent blood loss or transfusion, hemolytic anemia, iron deficiency anemia)
- In these situations, use glucose criteria exclusively for diagnosis
The Test Must Be Properly Performed
The A1c must be performed in a CLIA-certified laboratory using an NGSP-certified method standardized to the DCCT assay 1, 2, 3. Point-of-care A1c testing is not recommended for diagnostic purposes due to lack of mandated proficiency testing 2.
Clinical Bottom Line
Until you obtain confirmatory testing, this patient remains in the prediabetes category (A1c 5.7–6.4%) 1, 2, 4. Do not make a diabetes diagnosis, do not start diabetes medications, and do not document diabetes in the medical record based on this single value. Schedule repeat A1c testing promptly—not in 3 months, but soon to rule out laboratory error 2.
The single A1c of 6.5% should prompt you to retest, but it does not by itself establish the diagnosis. This approach protects the patient from premature labeling with a chronic disease diagnosis and its associated implications for insurance, employment, and psychological well-being, while ensuring that true diabetes is not missed through appropriate follow-up.