Current Diagnostic Levels of HbA1c
An HbA1c level of ≥6.5% (≥48 mmol/mol) is diagnostic for diabetes, while levels of 5.7% to 6.4% (39-46 mmol/mol) indicate prediabetes or high risk for developing diabetes. 1, 2
Diabetes Diagnosis
The diagnostic threshold is HbA1c ≥6.5% (≥48 mmol/mol) using an NGSP-certified method performed in an accredited laboratory. 3, 1, 2
Confirmation is required through either a repeat HbA1c test on a different sample or a glucose-based test (fasting glucose or OGTT), unless the patient presents with unequivocal hyperglycemia and classic symptoms. 1, 2
For borderline cases (HbA1c 6.5-6.9%), confirm with a fasting blood glucose level of ≥126 mg/dL (≥7.0 mmol/L) to account for racial/ethnic variations and measurement variability. 3
Prediabetes Diagnosis
HbA1c levels of 5.7% to 6.4% (39-46 mmol/mol) identify individuals at high risk for developing diabetes who would benefit from prevention interventions. 3, 1, 2
This range was established by the American Diabetes Association, though some international guidelines use 6.0-6.4% as the prediabetes threshold. 3
Critical Testing Requirements
Only use NGSP-certified methods standardized to the Diabetes Control and Complications Trial (DCCT) assay. 3, 2
Testing must be performed in CLIA-certified laboratories of moderate complexity or higher. 2
Point-of-care HbA1c devices should not be used for diagnosis unless specifically FDA-approved for diagnostic purposes and performed in appropriately certified laboratories. 2
Important Caveats and Limitations
HbA1c cannot be used for diagnosis in specific conditions:
- Hemoglobinopathies (sickle cell disease, thalassemia) that interfere with measurement. 2
- Pregnancy or gestational diabetes screening. 3, 2
- Conditions with increased red blood cell turnover (hemolytic anemia, recent blood loss, iron deficiency anemia). 3, 2
- Chronic kidney disease, which may alter HbA1c measurement. 3
Measurement considerations:
- Sequential HbA1c values within 0.5% do not statistically differ unless the assay coefficient of variation is <3% (ideally <2%). 3
- Marked discordance between HbA1c and plasma glucose should raise suspicion for assay interference and prompt use of glucose-based criteria instead. 2
- HbA1c at 6.5% identifies only about 30% of diabetes cases that would be diagnosed collectively using HbA1c, fasting glucose, or 2-hour glucose testing, meaning it has lower sensitivity than glucose-based criteria. 2
Historical Context
Prior to 2009-2010, HbA1c was not recommended for diagnosis. The 2003 guidelines explicitly stated HbA1c should not be used diagnostically due to lack of standardization and difficulty assigning appropriate cutpoints. 3
The shift to using HbA1c for diagnosis became possible through improved assay standardization via the NGSP and new data demonstrating associations between HbA1c concentrations and risk for retinopathy. 3