HbA1c Calculator
Medically reviewed by Dr. Chhaya Makhija, MD, DipABLM — September 2026
Convert between HbA1c (%) and estimated average glucose (eAG) using the validated ADAG formula. Enter your A1c to see average glucose — or enter average glucose to estimate A1c. Supports mg/dL and mmol/L. Educational reference only.
🔬 A1c ↔ Average Glucose Converter
Result
mg/dL
mmol/L
%
Category
An A1c result is a percentage, and a percentage of something most people never have explained to them. It is not an average of your meter readings, and it is not a blood sugar level. It is the proportion of your haemoglobin that has glucose chemically stuck to it — which turns out to be a reasonable proxy for how much glucose those cells have been swimming in for the past few months.
That indirectness is why A1c behaves in ways that surprise people. It can be perfectly normal in someone whose glucose swings between 50 and 300 every day. It can read high in someone with iron deficiency whose glucose is fine. And it can disagree sharply with a CGM average that seems like it should be measuring the same thing.
The converter below translates A1c into an average glucose figure using the ADAG equation, in both directions and in either unit. Below it you will find the reference table, the same result expressed in mmol/mol for labs outside the US, and — most importantly — the conditions under which the number should not be trusted at face value.
A1c Reference Table (ADA 2024)
| HbA1c (%) | eAG (mg/dL) | eAG (mmol/L) | ADA Category |
|---|---|---|---|
| 5.0 | 97 | 5.4 | Normal |
| 5.7 | 117 | 6.5 | Prediabetes threshold (lower) |
| 6.4 | 137 | 7.6 | Prediabetes threshold (upper) |
| 6.5 | 140 | 7.8 | Diabetes diagnosis threshold |
| 7.0 | 154 | 8.6 | ADA target (most adults) |
| 7.5 | 169 | 9.4 | Above target — reassess regimen |
| 8.0 | 183 | 10.2 | Uncontrolled — action needed |
| 9.0 | 212 | 11.8 | Significantly elevated |
| 10.0 | 240 | 13.4 | High risk — urgent review |
Source: ADA Standards of Care in Diabetes — 2026; Nathan DM et al. ADAG Study. Diabetes Care. 2008.
ADAG Formula
The ADAG (A1c-Derived Average Glucose) formula was derived from a multi-center study correlating lab A1c with continuous glucose monitor data across ~500 patients. It replaced earlier approximations and the ADA endorses it. The relationship is linear — each 1% A1c change ≈ 28.7 mg/dL (1.6 mmol/L) in average glucose.
A1c has known limitations. Results can be falsified by hemoglobin variants (sickle cell trait, thalassemia), hemolytic anemia, iron deficiency, and high-dose vitamin C. In these conditions, fructosamine or continuous glucose monitoring may be more reliable. Discuss with your provider if your A1c result seems inconsistent with your meter readings.
What HbA1c Actually Measures
HbA1c (glycated hemoglobin) is the share of your hemoglobin — the oxygen-carrying protein in red blood cells — that has glucose bound to it. Because red blood cells live about three months, A1c reflects your average glucose exposure over roughly the previous 8–12 weeks, weighted toward the most recent four. Unlike a finger-stick, it can't be "gamed" by a few good days before the appointment.
That long window is both its strength and its weakness: A1c is an excellent measure of overall trend, but it hides the day-to-day swings — the lows and post-meal spikes — that two people sharing the same 7.0% can experience very differently. To see your A1c as a familiar glucose number, use the converter above — and if you wear a CGM, compare it with your Glucose Management Indicator (GMI).
How Often Should HbA1c Be Tested?
The ADA's general guidance:
- Twice a year if you are meeting your treatment goals and glucose is stable.
- Every 3 months (quarterly) if your therapy has changed or you are not yet at goal.
Because the test averages months of data, checking it more often than about every three months rarely adds useful information — there simply hasn't been enough time for a meaningful change to register.
A1c in mmol/mol — the IFCC Unit
If your lab report says 53 rather than 7.0, nothing is wrong. Most of the world outside the United States reports A1c in mmol/mol using the IFCC standard, while the US uses the DCCT/NGSP percentage. They measure the same thing on different scales.
Example: an A1c of 7.0% → (7.0 − 2.15) × 10.929 = 53 mmol/mol.
| NGSP (%) | IFCC (mmol/mol) | eAG (mg/dL) | eAG (mmol/L) |
|---|---|---|---|
| 5.0 | 31 | 97 | 5.4 |
| 5.7 | 39 | 117 | 6.5 |
| 6.5 | 48 | 140 | 7.8 |
| 7.0 | 53 | 154 | 8.6 |
| 7.5 | 58 | 169 | 9.4 |
| 8.0 | 64 | 183 | 10.2 |
| 9.0 | 75 | 212 | 11.8 |
| 10.0 | 86 | 240 | 13.4 |
A useful anchor: the diagnostic threshold of 6.5% is 48 mmol/mol, and the common adult target of 7.0% is 53 mmol/mol.
When A1c Misleads — and Which Way
A1c depends on two things: how much glucose is around, and how long red blood cells survive to be exposed to it. Anything that changes red cell lifespan changes the result without glucose having moved at all — and knowing the direction of the error is what makes the information useful.
| Condition | Effect on A1c | Mechanism |
|---|---|---|
| Iron-deficiency anemia | Falsely high | Red cells survive longer, accumulating more glucose |
| Splenectomy | Falsely high | Reduced red cell clearance extends lifespan |
| Haemolytic anemia | Falsely low | Cells destroyed before they can glycate |
| Recent blood loss or transfusion | Falsely low | Young or donor cells with little glucose exposure |
| Chronic kidney disease / dialysis | Falsely low | Shortened red cell survival; erythropoietin therapy |
| Pregnancy (2nd–3rd trimester) | Falsely low | Increased red cell turnover and plasma volume |
| Haemoglobin variants (HbS, HbC, HbE) | Assay-dependent | Interferes with measurement rather than biology |
| High-dose vitamin C or E | Falsely low | Inhibits the glycation reaction |
Directions are typical, not guaranteed — the magnitude depends on severity and on which assay the lab runs.
The practical rule: when A1c and your meter or CGM tell different stories, believe the one with more data points. Any of the above can distort a single A1c summarising three months; a fortnight of CGM readings cannot be distorted in the same way. If the two disagree meaningfully, that discrepancy is itself worth raising with your care team, because fructosamine or glycated albumin may give a cleaner answer.
Why Your CGM Number and Your Lab A1c Disagree
If you wear a CGM, your app shows a Glucose Management Indicator — a figure that looks like an A1c and is often assumed to be a prediction of one. It is not. GMI is calculated from your CGM's mean glucose using a different equation, over a different window.
| Lab A1c | GMI (from CGM) | |
|---|---|---|
| What it measures | Glycated haemoglobin | Mean sensor glucose |
| Time window | ~8–12 weeks, recency-weighted | Typically the last 14 days |
| Affected by red cell lifespan | Yes | No |
| Affected by sensor accuracy | No | Yes |
Because they measure different things over different periods, a gap of up to roughly 0.5 percentage points in either direction is ordinary and not evidence that either is broken. A consistently large gap, however, usually means one of two things: your red cell turnover differs from the population average that A1c assumes, or your CGM is reading with a systematic bias. Both are worth investigating — and neither is fixed by trusting whichever number you prefer. Our GMI vs A1c guide works through how to interpret a persistent gap.
A1c summarises months of glucose but says nothing about the swings inside that average. Two people with the same 7.0% can spend very different amounts of the day in range, which is why time in range is now reported alongside it. If you are converting readings between units before entering them here, the blood glucose unit converter handles mg/dL and mmol/L.
Frequently Asked Questions
What is a good HbA1c level?
The ADA target for most non-pregnant adults with diabetes is below 7%. Some people aim for below 6.5%, while a looser goal below 8% may suit older adults or those at high risk of hypoglycemia. Below 5.7% is considered normal.
How do I convert my HbA1c to average glucose?
Use the ADAG formula: eAG (mg/dL) = 28.7 × A1c − 46.7. An A1c of 7% is about 154 mg/dL (8.6 mmol/L). The calculator above does this in both directions.
How often should I get an HbA1c test?
Generally twice a year if you are at goal and stable, or every 3 months if your treatment changed or you are not yet at target. More frequent testing rarely helps because A1c reflects months of data.
Can anything make an HbA1c result inaccurate?
Yes. Hemoglobin variants (such as sickle cell trait), hemolytic or iron-deficiency anemia, recent blood loss, pregnancy, and high-dose vitamin C can skew A1c. If your result doesn't match your meter, ask your provider about CGM or fructosamine testing.
What is 7% A1c in mmol/mol?
7.0% equals 53 mmol/mol. The conversion is IFCC = (NGSP% − 2.15) × 10.929. Two other anchors worth memorising: the diagnostic threshold of 6.5% is 48 mmol/mol, and 8.0% is 64 mmol/mol. Most labs outside the United States report in mmol/mol.
Does a falsely high A1c mean I have diabetes?
Not necessarily. Iron-deficiency anemia and splenectomy both raise A1c without glucose being elevated, because red cells survive longer and accumulate more glucose. This matters most near the 6.5% diagnostic threshold, where a false elevation could produce a diagnosis that a fasting glucose or oral glucose tolerance test would not support. Diagnosis normally requires two abnormal results, ideally by different methods.
Why is my CGM's GMI different from my lab A1c?
They measure different things over different windows. GMI comes from your sensor's mean glucose over about 14 days; A1c reflects glycated haemoglobin over 8–12 weeks, and red cell lifespan influences it. A difference of up to roughly 0.5 percentage points is normal. A persistently larger gap suggests either unusual red cell turnover or sensor bias, and is worth discussing with your care team.
Can A1c tell me my fasting blood sugar?
No — and this is a common misreading. The ADAG equation estimates your average glucose across the whole day and night, not your fasting value. Fasting glucose is typically lower than the 24-hour average in people whose post-meal excursions are large, and can be higher than average in people with a strong dawn phenomenon. Only a fasting measurement gives you a fasting number.
Sources
- Nathan DM et al. "Translating the A1C assay into estimated average glucose values." Diabetes Care. 2008;31(8):1473–1478. Link
- American Diabetes Association. Standards of Care in Diabetes — 2026. Section 6. Link
Written by Ryan Mitchell · Last updated: September 2026