A1C Diagnostic Threshold for Diabetes
An A1C level of ≥6.5% (≥48 mmol/mol) is the diagnostic threshold for diabetes, and this test must be performed using a method certified by the National Glycohemoglobin Standardization Program (NGSP) and standardized to the Diabetes Control and Complications Trial (DCCT) assay. 1, 2
Diagnostic Algorithm
When using A1C to diagnose diabetes, follow this approach:
- If A1C ≥6.5%: Confirm diagnosis with a repeat A1C test on a different sample OR a fasting plasma glucose test 2
- If A1C 5.7-6.4%: Diagnose prediabetes and initiate prevention strategies 2, 3
- If A1C <5.7%: Diabetes is not present by A1C criteria 3
In the absence of unequivocal hyperglycemia (random glucose ≥200 mg/dL with classic symptoms), two abnormal test results are required for diagnosis - either two separate A1C measurements ≥6.5%, or one A1C ≥6.5% plus one fasting glucose ≥126 mg/dL 1, 2
Rationale for the 6.5% Threshold
The 6.5% cutpoint was established based on its association with an inflection point for retinopathy prevalence, similar to the glucose-based diagnostic thresholds 1. Research confirms that A1C ≥6.5% and fasting glucose ≥7.0 mmol/L (126 mg/dL) are associated with significantly increased risk of incident retinopathy over 5 years 4.
Critical Situations Where A1C Should NOT Be Used
Do not use A1C for diagnosis in the following conditions - use plasma glucose criteria exclusively: 2
- Sickle cell disease or sickle cell trait (may lower A1C by ~0.3%) 1
- Pregnancy 1, 2
- Glucose-6-phosphate dehydrogenase deficiency (may lower A1C by 0.7-0.8%) 1
- Hemodialysis 1, 2
- Recent blood loss or transfusion 2
- Erythropoietin therapy 2
- Iron-deficiency anemia 1, 2
- HIV treated with certain antiretroviral drugs 1, 2
- Hemolytic anemias 1
Test Performance Characteristics
A1C ≥6.5% demonstrates high specificity (97-98%) but moderate sensitivity (47-67%) for detecting diabetes compared to glucose-based criteria 2. This means:
- A1C will identify approximately one-third fewer cases of undiagnosed diabetes than fasting glucose ≥126 mg/dL 1
- When A1C is elevated ≥6.5%, it is highly likely to confirm true diabetes 2
- Some individuals with diabetes by glucose criteria will be missed by A1C alone 5, 6
Important Clinical Caveats
Marked discordance between A1C and plasma glucose levels should raise suspicion for:
- Hemoglobin variants causing assay interference 1
- Conditions affecting red blood cell turnover 1
- Racial/ethnic differences in A1C levels independent of glycemia (African Americans may have A1C levels 0.3% higher than non-Hispanic whites at similar glucose levels) 1
For patients with hemoglobin variants but normal red cell turnover (such as sickle cell trait), use an A1C assay without interference from abnormal hemoglobins 1. An updated list of A1C assays with interferences is available at www.ngsp.org/interf.asp 1.
Advantages of A1C Testing
A1C offers several practical advantages over glucose-based testing 1, 3:
- No fasting required 1, 3
- Greater preanalytical stability 1, 3
- Less day-to-day variability during stress and illness 1
- Reflects average glucose over 60-90 days 3
- Already familiar to clinicians as a marker of glycemic control 1
Point-of-Care Testing Limitation
Point-of-care A1C assays are not sufficiently accurate for diagnostic purposes - only laboratory-based NGSP-certified methods should be used 1, 3