The A1C to eAG formula, explained

The short version

Two numbers describe the same underlying biology. A1C is the percentage of hemoglobin in your blood that has glucose permanently attached. eAG (estimated average glucose) restates that percentage in the everyday units your glucose meter uses — mg/dL in the US, mmol/L in most of the rest of the world. One formula moves between them:

Both directions use the same straight-line relationship — only algebraically rearranged. Everything after this section explains where that line comes from and how to use it responsibly.

Where the formula comes from: the ADAG study

The constants 28.7 and 46.7 are not arbitrary; they're the slope and intercept of a regression line published in 2008 by the A1c-Derived Average Glucose (ADAG) study group, led by Dr. David Nathan and funded in part by the American Diabetes Association, EASD, and IDF. The study enrolled 507 participants — including people with type 1 diabetes, type 2 diabetes, and no diabetes — across ten centers on three continents.

Each participant wore a continuous glucose monitor and performed frequent self-monitoring finger-sticks for roughly three months, generating thousands of glucose data points per person. Those readings were averaged and statistically weighted, then compared against a central-laboratory A1C measurement drawn at the end of the same window. The result was a linear relationship strong enough (a correlation coefficient of about 0.92) that the ADA, EASD, and IDF jointly endorsed reporting eAG alongside A1C on lab results, so patients could see their long-term control in the same units as their daily readings.

A worked example

Say a lab report shows an A1C of 7.2%. Plugging that into the mg/dL formula:

In mmol/L, the same 7.2% works out to 1.59 × 7.2 − 2.59 ≈ 8.8 mmol/L. Both numbers describe the same three-month average — a person with this A1C has been living, on average, with a blood sugar in the high 150s to low 160s (mg/dL), even though any single reading that day might have been 90 after a fast or 220 after a large meal.

NGSP percent vs. IFCC mmol/mol: two different unit systems

Separately from eAG, there's a second conversion that trips people up: A1C itself is reported in two different unit systems depending on the country. The NGSP (National Glycohemoglobin Standardization Program) unit is the familiar percentage used across the US, tied historically to the DCCT and UKPDS trials. The IFCC (International Federation of Clinical Chemistry) unit, used in the UK, much of Europe, and Australia since around 2011, expresses the same measurement in millimoles of glycated hemoglobin per mole of total hemoglobin (mmol/mol) using a more analytically specific reference method.

The two scales are linearly related by a master equation maintained jointly by the NGSP and IFCC networks:

A worked example: an NGSP result of 6.5% converts to (6.5 − 2.15) × 10.929 ≈ 47.5, rounded to the commonly quoted 48 mmol/mol — the UK's NICE diagnostic threshold for diabetes, equivalent to the ADA's 6.5% threshold. A 7.0% reads as roughly 53 mmol/mol. If you're comparing an NHS or European lab report to an American one, this is the conversion you need — it's a completely separate calculation from the eAG formula above, even though both are often shown on the same chart.

Quick NGSP ↔ IFCC reference points

• 5.7% ≈ 39 mmol/mol (prediabetes threshold, ADA)
• 6.0% ≈ 42 mmol/mol
• 6.5% ≈ 48 mmol/mol (diabetes threshold, ADA & NICE)
• 7.0% ≈ 53 mmol/mol (common treatment target)
• 8.0% ≈ 64 mmol/mol
• 9.0% ≈ 75 mmol/mol

Why the formula won't match your meter exactly

The ADAG relationship is a population average, not a personal guarantee. Two people can share an identical true average glucose and still land on visibly different A1C values, for a few well-documented reasons:

• Red blood cell lifespan. A1C only forms while a red blood cell is circulating. Anything that shortens that lifespan — hemolytic anemia, certain medications, chronic kidney disease, recent significant blood loss — gives glucose less time to bind, and A1C reads artificially low relative to true average glucose. Conditions that lengthen red cell survival, such as iron-deficiency anemia or a splenectomy, can push it artificially high.
• Individual glycation rate. Independent of red cell lifespan, some people's hemoglobin simply binds glucose faster or slower than the population average — sometimes informally called being a "high glycator" or "low glycator." Research suggests this can shift A1C by the equivalent of several tenths of a percentage point at the same average glucose.
• Hemoglobin variants. Sickle cell trait, thalassemia, and other hemoglobinopathies can interfere with certain A1C assay methods, producing results that don't reflect glucose control at all. Labs typically flag when an alternative test is needed.
• Recent transfusion or pregnancy. A transfusion introduces donor red cells with a different glycation history, and the physiological changes of pregnancy alter red cell turnover — both are reasons A1C is considered unreliable or is not used at all in those situations.
• Sampling difference. A meter average is built from however many finger-sticks or CGM readings you happened to take; A1C reflects a continuous, unbroken 24-hour record including sleep. The two are answering slightly different questions even when everything else is normal.

How labs actually measure A1C

The formula only matters because a physical lab test produces the A1C percentage in the first place. Common methods include high-performance liquid chromatography (HPLC), which separates hemoglobin fractions by their electrical charge; immunoassays, which use antibodies that specifically recognize the glycated portion of the protein; boronate affinity chromatography, which binds glucose-modified hemoglobin directly; and enzymatic assays increasingly used in point-of-care devices. All NGSP-certified methods are required to correlate tightly with the original DCCT reference method so that a result from one certified lab means the same thing as a result from another.

A brief history of standardization

Before the 1990s, A1C assays varied enough between labs that a result from one hospital wasn't reliably comparable to another. The landmark Diabetes Control and Complications Trial (DCCT, 1983–1993) and UK Prospective Diabetes Study (UKPDS, 1977–1997) demonstrated that A1C predicted long-term complication risk, which created urgent pressure to standardize the measurement itself. The NGSP was formed in the mid-1990s to align US labs to the DCCT reference method. The IFCC later developed an even more chemically specific reference method, adopted as the primary reporting standard in the UK and much of Europe from 2011 onward, which is why two valid unit systems now coexist worldwide.