What are the sources for A1C calculations?

The equations are from Nathan and colleagues, the A1C-Derived Average Glucose Study Group, Translating the A1C Assay Into Estimated Average Glucose Values, Diabetes Care, August 2008, https://web.archive.org/web/20120219165011/care.diabetesjournals.org/content/31/8/1473.full. The lines are eAG (mg/dL) = 28.7 × A1C − 46.7 and eAG (mmol/L) = 1.59 × A1C − 2.59.

A1C Calculator

ADAG conversion between an A1C percent and estimated average glucose. An estimate, not a diagnosis.

Smart Tips

• From A1C, mg/dL is 28.7 × A1C − 46.7. mmol/L is 1.59 × A1C − 2.59.

• An A1C of 7% is 154.2 mg/dL and 8.54 mmol/L on those two lines.

• The two lines are the published regressions. The page does not convert mg/dL to mmol/L by dividing by 18.

• Going backward, 154.2 mg/dL returns 7%, and 8.54 mmol/L returns 7%.

• If the line falls below zero, the page says so. That is not a glucose you can use.

• There is no diabetes label and no medication dose on this page.

The Formula

eAG (mg/dL) = 28.7 × A1C − 46.7
eAG (mmol/L) = 1.59 × A1C − 2.59

28.7 mg/dL slope
ADAG linear regression

46.7 mg/dL intercept
Subtracted after the slope

1.59 mmol/L slope
Published SI form of the same study

2.59 mmol/L intercept
Published with the SI slope

The paper reports an R² of 0.84 and a prediction-error standard deviation of 15.7 mg/dL (0.87 mmol/L). The result is an estimate around that scatter, not a meter average.

Key Insights

One conversion

The page turns A1C into estimated average glucose, or the reverse. It does not interpret the number.

Two published lines

mg/dL and mmol/L each use the equation printed in the ADAG paper. They are not forced through a divide-by-18 step.

About 15 mg/dL of scatter

The study’s prediction-error standard deviation is 15.7 mg/dL. A single A1C does not pin glucose to the nearest point.

No label and no dose

Nothing on this page says diabetes, prediabetes, or a medication amount. Those decisions are not this equation.

Medical Disclaimer

Estimate only: This page is an estimate, not a diagnosis, a lab test, or a treatment plan.

No label and no dose: This page does not say whether someone has diabetes, and it does not recommend a medication dose or a treatment change.

Frequently Asked Questions

Estimated average glucose in mg/dL is 28.7 times the A1C percent, minus 46.7. In mmol/L it is 1.59 times the A1C percent, minus 2.59. Both lines are the ADAG regressions.

28.7 × 7 − 46.7 is 154.2 mg/dL. 1.59 × 7 − 2.59 is 8.54 mmol/L. Running either number backward returns 7 percent. The page does not attach a diagnosis to 7.

The paper published both regressions. This page uses each one as printed. Dividing the mg/dL line by 18 is a close unit conversion, not the mmol/L equation the study stated.

The ADAG paper reports R² = 0.84 and a prediction-error standard deviation of 15.7 mg/dL (0.87 mmol/L). The number is an estimate with that scatter. It is not your meter’s average.

No. It does not print a diabetes or prediabetes label for any input, and it does not suggest a dose. The result is only the ADAG conversion.

The equations are from Nathan and colleagues, the A1C-Derived Average Glucose Study Group, “Translating the A1C Assay Into Estimated Average Glucose Values,” Diabetes Care, August 2008, Diabetes Care 2008;31:1473-1478. The lines are eAG (mg/dL) = 28.7 × A1C − 46.7 and eAG (mmol/L) = 1.59 × A1C − 2.59.

Responsive Ad Space

Estimated average glucose

Enter values to calculate

mg/dL

mmol/L

Note:ADAG only: 28.7 × A1C − 46.7 mg/dL, and 1.59 × A1C − 2.59 mmol/L. The paper’s prediction error is about 15.7 mg/dL. An estimate, not a diagnosis. No diabetes label and no dose.