A1C Calculator

The formula and where it comes from

The A1C-Derived Average Glucose (ADAG) study, published in Diabetes Care in 2008, followed 507 adults (268 with type 1 diabetes, 159 with type 2 and 80 without diabetes) for three months. Each participant contributed about 2,700 glucose readings, mostly from continuous monitoring plus fingersticks, and researchers compared that measured average with lab A1C. The best-fitting line was:

eAG (mg/dL) = 28.7 × A1C − 46.7

eAG (mmol/L) = 1.59 × A1C − 2.59

The correlation was strong (R² = 0.84). This calculator uses the first equation for mg/dL and the second, ADAG's own mmol/L equation, for mmol/L, so its mmol/L column can differ by 0.1 from printed charts that divide a rounded mg/dL value by 18. In the reverse direction it solves A1C = (eAG + 46.7) ÷ 28.7, rounds to one decimal place, and converts the glucose you entered to mmol/L by dividing by 18.0182.

A1C to eAG conversion chart

Values below are produced by the same formula the calculator uses.

Each 1-point rise in A1C adds 28.7 mg/dL (about 1.6 mmol/L) to the estimated average. Published charts often round to whole numbers, so you may see 126 mg/dL for 6% and 183 mg/dL for 8%.

How precise is a single eAG number?

Less precise than one decimal place suggests. The ADAG paper reports 95% confidence ranges around each estimate: at an A1C of 7%, the average glucose of individual participants ranged from about 123 to 185 mg/dL; at 8%, from 147 to 217 mg/dL. Two people with the same A1C can therefore have meaningfully different day-to-day glucose. The study also excluded children and pregnant women, had few participants of African descent and only included people with fairly stable glucose control, so the equation is least certain outside that group.

Why your meter or CGM average doesn't match your A1C

• Sampling. A fingerstick meter only sees the moments you test, usually before meals, and misses overnight and post-meal values.
• Red blood cell lifespan. A1C depends on how long red cells circulate. NIDDK notes that recent blood loss, sickle cell disease and iron-deficiency anemia can push A1C up or down regardless of glucose.
• Hemoglobin variants. Some variants more common in people of African, Mediterranean or Southeast Asian ancestry can make standard A1C assays read falsely high or low.
• Other conditions. CDC lists kidney failure, liver disease, pregnancy, transfusions and certain medicines, including opioids and some HIV drugs, as factors that can affect A1C accuracy.

eAG vs GMI on a CGM report

CGM software usually shows a Glucose Management Indicator (GMI) rather than eAG. GMI runs the conversion the other way, predicting A1C from CGM mean glucose with a different equation: GMI (%) = 3.31 + 0.02392 × mean glucose (mg/dL). It was renamed from "estimated A1C" because patients were confused when it didn't match their lab result. In the validation data, 28% of people had a GMI at least 0.5 points away from their lab A1C, and only 19% had identical values. The authors recommend at least 10, preferably 14 or more, days of CGM data.

Because the formulas differ, a CGM mean of 160 mg/dL gives an A1C of 7.2% in this calculator (ADAG) but a GMI of about 7.1%. Neither is wrong; they are different regressions from different studies.

A1C in mmol/mol (IFCC units)

Labs in the UK, much of Europe and Australia report A1C in mmol/mol (IFCC units) rather than percent (NGSP units). The two scales are linked by the NGSP master equation: NGSP (%) = 0.09148 × IFCC (mmol/mol) + 2.152, so IFCC = (NGSP − 2.152) ÷ 0.09148. This calculator expects A1C in percent; convert a mmol/mol result first.

Do not confuse A1C in mmol/mol with glucose in mmol/L. An A1C of 53 mmol/mol is 7.0%, while an average glucose of 8.5 mmol/L is the eAG for that same A1C.

What the A1C ranges mean

NIDDK and CDC use the same diagnostic categories for lab A1C:

For many adults with diabetes the general goal is an A1C below 7%, but targets are individualized based on age, other conditions and the risk of hypoglycemia. CDC notes that most people with diabetes have A1C tested at least twice a year and that people with prediabetes are usually retested every one to two years. A converted eAG cannot be used to place yourself in these categories; only a lab test, confirmed as your clinician decides, can. Ask your doctor what target is right for you.