A1C to Average Glucose Converter
Convert A1C to estimated average glucose (eAG) in mg/dL or mmol/L and back, with the ADA reference chart. Runs entirely in your browser.
About this tool
A1C and average glucose measure the same thing from two sides. A1C is the percentage of hemoglobin that has glucose attached — a reading taken in a lab, on a blood draw. Estimated average glucose (eAG) is what that same number would look like on a glucose meter, averaged over the roughly 2–3 months that red blood cells live. This tool converts between the two in both directions, using the formula from the ADAG study that the American Diabetes Association adopted: eAG (mg/dL) = 28.7 × A1C − 46.7.
The conversion is linear, but the numbers do not feel linear: moving from A1C 6% to 7% adds 29 mg/dL, while moving from 9% to 10% adds another 29 on top of a much higher base. That is why the chart below lists 19 rows instead of one formula — the step from 5% to 6% and the step from 9% to 10% look small on paper and are not small in daily readings.
The three colored ranges come straight from the ADA’s diagnostic reference values: below 5.7% is below the prediabetes threshold, 5.7–6.4% is the prediabetes range, and 6.5% or above is the diabetes range. They are included so the number has context — they are a restatement of published thresholds, not this tool making a judgment about you.
How the chart decides which row to highlight
Your A1C is rounded to the nearest 0.5 and the matching row is marked — 7.2 marks the 7.0 row. Values outside the 5.0–14.0 range clamp to the nearest end row instead of leaving the chart blank.
The highlight is arithmetic rather than a lookup of your exact value: the input is divided by the step, rounded, and clamped into the table. So anything from 6.75 up to but not including 7.25 marks the 7.0 row, and the highlight is there for orientation rather than as a second calculation.
The clamping matters more than it looks. A value below 5.0 or above 14.0 — both possible, since the tool accepts A1C from 3 to 20 — marks the 5.0 or the 14.0 row rather than returning nothing. The chart stays useful as a ruler even when your number sits off its end.
Why the mmol/L figure is derived, not converted separately
mg/dL is the only stored value; mmol/L is always that number divided by 18.018. Because one reading is derived from the other, the two can never drift apart or disagree about the same sample.
Glucose is measured two ways: mass concentration in mg/dL, used in the United States, and molar concentration in mmol/L, used in much of Europe and Asia. The two are related by the molar mass of glucose, so the conversion is a single division by 18.018 — the same factor the ADA tables use.
The reason it is a derivation rather than a second conversion is consistency. If mmol/L ran along its own path, a rounding difference could make a value converted from A1C and a value converted from a glucose reading disagree in the third decimal place. Deriving every output from one canonical mg/dL number removes that possibility.
The exact molar mass of glucose would give a factor near 18.016; the tool uses 18.018 to stay consistent with the published ADA tables rather than to chase the last digit.
Why the A1C and glucose input limits do not line up
A1C is accepted from 3 to 20%, glucose from 40 to 700 mg/dL — and those ranges are not the same span. A1C 3% converts to 39.4 mg/dL, just under the glucose floor, while 700 mg/dL converts back to about 26%.
Each side has its own guard, set independently: the A1C field rejects anything outside 3–20 because values beyond that are almost certainly a typo, and the glucose field does the same at 40–700 mg/dL. Because the two limits are not exact images of each other under the formula, the outputs you can reach differ slightly depending on which field you type into.
In practice: type A1C 3% and the glucose shown is 39.4 mg/dL, fractionally below the 40 you could have entered directly. Type 700 mg/dL and the A1C shown is about 26%, above the 20 you could have entered. Neither is a bug — each field guards its own input, and the conversion between them is one line of algebra shared by both.
The same A1C can mean different average glucose
The eAG on this page is an estimate, not a measurement. It comes from a population regression, and A1C itself shifts with red-cell lifespan, haemoglobin variants, anaemia and kidney disease — so 7.0% does not mean the same average glucose in two people.
The conversion comes from the ADAG study (Nathan et al., 2008): ten laboratories measuring the same people, giving eAG = 28.7 × A1C − 46.7. It is a regression result — it describes the average relationship across a population, not a precise correspondence for any one person. The scatter around that line is real, which is why the ADA calls it estimated A1C-derived average glucose, with the emphasis on the word estimated.
Where the bias comes from: A1C is the fraction of haemoglobin that has been glycated, so anything that changes red-cell lifespan or the haemoglobin itself pushes the number away. Haemolysis, blood loss and iron-deficiency anaemia lower it; splenectomy and vitamin B12 deficiency raise it; haemoglobin variants (HbS or HbC trait, for example) can make certain assays read incorrectly. Kidney impairment shifts it too. In those situations the relationship between A1C and true average glucose is no longer this straight line.
So the practical use is this: treat A1C as a trend marker, where the direction over three months says more than any single reading, and use eAG to translate a percentage into the mg/dL most people think in — for communication, not as a value to reconcile against a meter. If your A1C and the average of your own readings disagree persistently, that is a signal worth raising with a clinician: it may mean one of the factors above is present, not that your meter is broken.
Frequently asked questions
Which formula does this use, and why that one?
The ADAG study formula (Nathan et al., 2008): eAG (mg/dL) = 28.7 × A1C − 46.7, with mmol/L derived by dividing by 18.018. It was established across 10 laboratories and 3 A1C assay methods with continuous glucose monitoring as ground truth, which is why the American Diabetes Association adopted it as the standard conversion. Older "rule of 28" shortcuts predate it and drift by several mg/dL.
My meter shows 110 mg/dL. Is that my A1C?
Not directly — a single meter reading is one moment, while A1C is a 2–3 month average. Only an averaged glucose (for example from a continuous glucose monitor report, or a lab-calculated eAG) should be converted to A1C. Converting a single fasting reading of 110 mg/dL gives an "A1C" near 5.5%, but your actual A1C also depends on what happens after meals and overnight.
Does the colored range mean I have diabetes?
No. The tool restates the ADA’s published reference thresholds so the number has context. A diagnosis is made by a clinician using confirmed lab measurements, sometimes repeated, together with the rest of your history. Use this page to understand what a lab report says, not to diagnose anything.
Does it send my numbers anywhere?
No. The conversion is one line of arithmetic in the page. There is no server, no request and no logging, and the tool keeps working with the network switched off — a practical way to verify the claim yourself.
Related tools
Work out your body mass index from metric or imperial measurements, see the WHO weight category and the healthy weight range for your height. Nothing you type leaves your browser.
Calculate your basal metabolic rate with three formulas side by side — Mifflin-St Jeor, Harris-Benedict and Katch-McArdle — in metric or imperial units. See why they disagree and which one fits you.
Total daily energy expenditure: BMR (Mifflin-St Jeor) times five standard activity levels, in one full table — plus calorie targets for losing, maintaining or gaining.
Find your five heart-rate training zones from age and resting heart rate. Four max-HR formulas and both the %HRmax and Karvonen methods, side by side.