A1c explained: the number behind the three-month average
One blood draw, no fasting required, and a single percentage that summarizes your blood sugar for the last three months. A1c is genuinely one of medicine's best tests — which makes the cases where it misleads all the more worth knowing.
What it actually measures
Glucose in your bloodstream sticks to hemoglobin, the oxygen-carrying protein inside red blood cells — permanently, for the life of the cell. A1c measures the percentage of your hemoglobin that has glucose stuck to it. Because red blood cells live about three months, that percentage works out to a weighted average of your blood sugar over roughly the last 8–12 weeks — weighted toward the most recent month, since younger cells outnumber older ones.
That's why A1c can't be gamed by eating clean for three days before the draw, and why it complements rather than replaces a fasting glucose: fasting glucose is a snapshot; A1c is the movie.
The cutoffs
| A1c | Category | Approx. average glucose |
|---|---|---|
| Below 5.7% | Normal | under ~117 mg/dL |
| 5.7% – 6.4% | Prediabetes | ~117–137 mg/dL |
| 6.5% or higher | Diabetes (confirmed by repeat or a second test) | ~140 mg/dL and up |
Two anchor points for the mapping: an A1c of 6% corresponds to an estimated average glucose around 126 mg/dL, and each additional percentage point adds roughly 28 mg/dL. For people already managing diabetes, the commonly cited treatment target is an A1c below 7% — but that target is individualized: tighter for some, deliberately looser for older adults or anyone at meaningful hypoglycemia risk. The right target is a conversation, not a lookup table.
Where the number quietly lies
A1c assumes your red blood cells live a normal three months. When they don't, the number drifts away from reality:
Reads falsely low — anything that shortens red-cell life or dilutes older cells with young ones: significant blood loss, hemolytic anemia, recent transfusion, some kidney disease, pregnancy (mid-to-late), and treatment with iron/B12 or erythropoietin during recovery.
Reads falsely high — conditions that let cells linger: iron-deficiency anemia (before treatment) and some other slow-turnover states.
Reads wrong in both directions — hemoglobin variants (sickle trait and others) can interfere with some lab methods; labs have workarounds, but only if someone knows to use them.
None of this makes A1c a bad test. It means a surprising A1c that doesn't match your home glucose readings is a data conflict worth raising, not a verdict to accept silently.
The trend is the treatment story
A single A1c diagnoses; a series tells you whether anything you're doing is working. Moving from 8.9% to 6.9% over a year — a real pattern in successful lifestyle-plus-metformin management — is a different fact than either number alone, and it's the fact that drives decisions. This is the same lesson as lab results generally: your own history is the reference range that matters.