Use of HbA1c for diagnoses of diabetes and prediabetes: comparison with diagnoses based on fasting and 2-hr glucose values and effects of gender, race, and age.
Level 3 - non-randomized controlled study
Cross-sectional diagnostic accuracy study using a national survey cohort
PubMed 24512556 · doi:10.1089/met.2013.0128
What was done
This study evaluated the diagnostic accuracy of standard HbA1c thresholds (6.5% for diabetes, 5.7% for prediabetes) compared against fasting plasma glucose (FPG) and 2-hour post-load glucose criteria. The analysis examined receiver operating characteristic (ROC) curves and demographic variations (age, gender, race) in 5,395 adults without known diabetes from the National Health and Nutrition Examination Survey (NHANES) 2005–2010.
What was found
When compared to diagnoses using both FPG and 2-hr glucose: - HbA1c cutoffs exhibited low sensitivity for diabetes (0.249) and prediabetes (0.354), alongside high specificity, yielding false-negative rates of 75.1% and 64.9%, respectively. - Misdiagnosis rates increased with age and were higher among non-Hispanic whites and Mexican Americans. - Combining HbA1c with FPG lowered the false-negative rate to 45.7% for diabetes and to 9.2% for prediabetes.
Why it matters
HbA1c measurements alone frequently fail to detect diabetes and prediabetes when compared to oral glucose tolerance testing, indicating that values below diagnostic cutoffs do not rule out dysglycemia.
Limits
The abstract does not provide exact numerical values for specificity or confidence intervals. Diagnoses were based on single-timepoint survey measurements rather than repeated clinical confirmatory tests, and longitudinal patient outcomes were not evaluated.
Cited by
- supports The HbA1c test misses over half of insulin resistance or metabolic health problems.