Calculatrice A1C - Hémoglobine A1C en Glycémie Moyenne

The HbA1c test measures the percentage of hemoglobin in red blood cells that has glucose permanently attached to it. Because red blood cells live for about three months, the result reflects average blood glucose over the preceding two to three months rather than glucose at a single moment. This calculator converts an A1C percentage into an estimated average glucose (eAG) value in mg/dL and mmol/L, and shows where that percentage sits against the American Diabetes Association diagnostic categories.

La formule

eAG (mg/dL) = 28.7 x A1C (%) - 46.7

• A1C = glycated hemoglobin percentage from a laboratory blood test
• eAG = estimated average glucose in mg/dL over roughly the preceding 2 to 3 months
• eAG (mmol/L) = eAG (mg/dL) divided by 18.0

ADAG (A1c-Derived Average Glucose) Study Group regression, published in Diabetes Care, 2008

How the calculation works

Glucose in the bloodstream attaches to hemoglobin inside red blood cells through a slow, largely irreversible chemical reaction called glycation. The higher the average glucose level, and the longer glucose stays elevated, the greater the fraction of hemoglobin that becomes glycated. Because red blood cells circulate for about 100 to 120 days, the A1C percentage measured today mostly reflects glucose exposure over the last 8 to 12 weeks, with the most recent 30 days contributing the largest share.

The ADAG study enrolled 507 participants across ten countries, each contributing roughly 2,700 glucose readings over 3 months, linking that data to laboratory A1C values and producing a linear regression. The correlation between mean glucose and A1C held strongly from about 4 percent to 13 percent, which is why a single linear formula is considered adequate across nearly the entire clinically relevant range. This calculator applies that regression directly: multiply the A1C percentage by 28.7, then subtract 46.7, to get eAG in mg/dL. Dividing by 18.0 converts the result to mmol/L, the unit used in most countries outside the United States.

How to read your result

The American Diabetes Association defines three categories using the A1C percentage alone. The eAG figure is provided as a more intuitive companion to the percentage and is not itself a separate diagnostic threshold.

Diagnosis, monitoring, and treatment targets

A diagnosis of diabetes based on A1C alone is usually confirmed with a repeat test on a separate day, unless the person has classic symptoms of hyperglycemia together with a random plasma glucose of 200 mg/dL or higher. For people already diagnosed with diabetes, the American Diabetes Association generally recommends an A1C target below 7 percent for most non-pregnant adults, while allowing a less stringent target, such as below 8 percent, for people with a limited life expectancy, extensive coexisting illness, or a history of severe hypoglycemia. Some younger, otherwise healthy people with a short duration of diabetes may be given a tighter target, such as below 6.5 percent, if it can be reached without significant hypoglycemia.

A1C is typically rechecked every 3 months when treatment has recently changed or targets are not being met, and every 6 months once glucose control is stable and at target. Because A1C is an average, it cannot distinguish between someone with consistently moderate glucose and someone who swings between hypoglycemia and hyperglycemia; for that reason, people using insulin or at risk of low blood glucose are also advised to track readings with a meter or continuous glucose monitor rather than relying on A1C alone.

A1C and continuous glucose monitoring metrics

Continuous glucose monitors report their own derived estimate of A1C, called the glucose management indicator (GMI), calculated from a separate regression built on CGM sensor data rather than the ADAG study's laboratory A1C and glucose data used for the eAG formula above. GMI and eAG both attempt to translate a glucose measure into an A1C-equivalent percentage, but they start from different inputs, so the two are not interchangeable: a person's CGM-derived GMI over a 14-day period can run meaningfully higher or lower than their next laboratory A1C, particularly when red blood cell turnover, and therefore the rate of glycation, is faster or slower than the population average both formulas assume.

International consensus guidance also tracks time-in-range (TIR), the percentage of CGM readings between 70 and 180 mg/dL (3.9 to 10.0 mmol/L) over a wear period of at least 14 days. TIR and A1C move together on average, roughly a 10 percentage point rise in TIR per 0.5 percentage point fall in A1C, but TIR captures what A1C cannot: two people can share an identical 7.0 percent A1C while one holds steady glucose most of the day and the other cycles between hypoglycemia and hyperglycemia that happen to average out to the same number.

• Time in range, 70 to 180 mg/dL (3.9 to 10.0 mmol/L): general target above 70 percent of readings for most adults with diabetes
• Time below range, under 70 mg/dL (3.9 mmol/L): general target below 4 percent of readings
• Time below range, under 54 mg/dL (3.0 mmol/L), marking clinically significant hypoglycemia: general target below 1 percent of readings
• Time above range, over 180 mg/dL (10.0 mmol/L): general target below 25 percent of readings

Limites

• A1C can be unreliable in people with conditions that shorten or lengthen red blood cell lifespan, including hemolytic anemia, recent significant blood loss, blood transfusion, and splenectomy; these conditions can push the result artificially low or high regardless of true average glucose.
• Hemoglobin variants such as sickle cell trait, HbC, and HbE can interfere with some A1C assay methods and produce an inaccurate result; a laboratory should be told about known hemoglobin variants so an appropriate assay is used.
• Iron deficiency anemia tends to raise A1C independent of glucose, while pregnancy and chronic kidney disease, including dialysis, can shift results in either direction; A1C is generally not used to diagnose diabetes during pregnancy.
• Average A1C can differ modestly across racial and ethnic groups at the same average glucose level, which is one reason A1C is interpreted alongside, not instead of, direct glucose testing when results are borderline.
• The eAG conversion is a population-derived average relationship, not a personal calibration; two people with an identical A1C can have somewhat different true average glucose, and day-to-day glucose variability is not captured by either number.

Questions fréquentes

Can A1C diagnose diabetes on its own?

Yes, an A1C of 6.5 percent or higher on two separate tests, or one test alongside classic symptoms and a high random glucose, meets the American Diabetes Association criteria for a diabetes diagnosis. A single borderline result is usually repeated to confirm it.

Why is my meter's average glucose different from my eAG?

eAG is a statistical estimate derived from a study population, not a direct measurement of your own glucose. Differences in glycation rate, testing patterns, and the time window each measure covers mean your device average and the calculated eAG will rarely match exactly.

How often should A1C be tested?

General guidance is every 3 months when adjusting treatment or not at target, and every 6 months once glucose is stable at goal. A clinician may test more or less often depending on individual circumstances.

Does fasting affect the A1C test?

No. Unlike a fasting glucose test, A1C reflects an average over months, so eating or drinking beforehand does not meaningfully change the result.

What is the glucose management indicator (GMI) and how does it differ from eAG?

GMI is a CGM's own estimate of A1C, calculated from sensor glucose data over a recent period, typically 14 days, using a regression built specifically for CGM data. eAG instead converts a measured lab A1C into an average glucose figure using the ADAG regression. Both approximate a similar relationship from opposite directions, so GMI and a subsequent lab A1C can differ for the same person.

What is a good time-in-range target?

General targets for most adults with diabetes are above 70 percent of CGM readings in range, below 4 percent below range, and below 25 percent above range, measured over at least 14 days of sensor wear. Targets are individualized and often relaxed for older adults or anyone prone to hypoglycemia.

Références

1. NIDDK. Diabetes Tests & Diagnosis.
2. CDC. Diabetes Testing.
3. American Diabetes Association. Understanding A1C.
4. NIDDK. The A1C Test and Race/Ethnicity.