A1C Blood Test: Chart, Normal Levels, and eAG Conversion

The A1C blood test measures the share of your hemoglobin that has been coated with glucose. Because red blood cells live about three months, A1C reflects your average blood sugar over the past two to three months — and you don't have to fast for it. That makes it the reference test for diagnosing and monitoring diabetes. This guide gives you the A1C chart, the eAG conversion into mg/dL, how to read a high A1C, and — the part most articles skip — what makes an A1C falsely high or low.

A note on names: A1C, HbA1c, hemoglobin A1C, and glycated hemoglobin all mean the same test. U.S. labs and the ADA say A1C; "HbA1c" is the form used in journals and abroad.

Key takeaways

• A1C reflects your average blood glucose over ~2–3 months and requires no fasting.¹
• A1C chart: normal below 5.7%, prediabetes 5.7–6.4%, diabetes 6.5% or above — usually confirmed with a second test.²
• Many people with diabetes are advised to keep A1C below 7%, but the goal is individualized.¹³
• eAG conversion: 6% ≈ 126 mg/dL · 7% ≈ 154 mg/dL · 8% ≈ 183 mg/dL average glucose.⁴
• In 2021, 38.4 million Americans (11.6%) had diabetes — 8.7 million adults undiagnosed — and 97.6 million adults, more than 1 in 3, had prediabetes.⁵
• The USPSTF recommends screening adults 35–70 with overweight or obesity, roughly every 3 years if normal.⁶
• A1C can be falsely high or low in iron deficiency, anemia, hemolysis, kidney disease, pregnancy, after transfusion or blood loss, and with some hemoglobin variants.⁷⁸

What is A1C?

Hemoglobin is the iron-rich protein inside your red blood cells that carries oxygen. When glucose circulates, some of it attaches to that hemoglobin permanently — glycation. The higher your blood sugar runs, the larger the glycated fraction. That fraction, as a percentage of your total hemoglobin, is your A1C.⁹ Because a red blood cell survives about 120 days, A1C is a memory of your blood sugar over the previous two to three months. A fasting glucose reading is a snapshot of one moment; A1C is the average of thousands.¹

In the United States A1C is reported in percent (%), standardized to the NGSP/DCCT reference method. Elsewhere the same result appears in mmol/mol under the international IFCC system — a different scale for the same molecule.¹⁰ Your U.S. report shows the percentage, and that is the number this guide uses.

Why is A1C tested?

Your primary care provider may order an A1C to screen for or diagnose type 2 diabetes and prediabetes, to monitor known diabetes — at least twice a year when control is stable, about every 3 months when it isn't or after a treatment change — and to judge whether a medication or lifestyle change is working.¹²⁸

Who should be screened? The U.S. Preventive Services Task Force recommends screening asymptomatic adults aged 35 to 70 with overweight or obesity (BMI ≥ 25, or ≥ 23 for Asian Americans) — a Grade B recommendation — roughly every 3 years when results are normal. Fasting glucose, A1C, and the OGTT are all acceptable.⁶ Anyone with prediabetes should be retested every year.²

A1C chart: normal levels and thresholds

These are the diagnostic categories used in the United States for non-pregnant adults. They come from NIDDK and match the ADA's Standards of Care, on a standardized, NGSP-certified assay:²¹¹

Good to know: a diagnosis is not made on one number — "usually, your doctor will use a second test to confirm you have diabetes."² And the treatment target is not the diagnostic threshold: some people with diabetes reduce their risk of complications by keeping A1C below 7%, but your own goal is set with your clinician — often tighter when newly diagnosed and young, deliberately looser in an older or fragile person for whom hypoglycemia is the bigger danger.¹³¹² A1C is not used to diagnose gestational diabetes.⁹

A1C to eAG conversion chart

A percentage is abstract; mg/dL is what you see on a glucose meter. The estimated average glucose (eAG) translates between them, using the ADAG study equation eAG (mg/dL) = (28.7 × A1C) − 46.7.⁴¹³

Each 1% of A1C is worth roughly 29 mg/dL of average glucose. Take the word estimated seriously: two people with an identical 7% A1C can have very different daily patterns — one steady, one swinging between highs and lows. That gap is why CGM and time in range have become standard companions to A1C.¹⁴

Interpreting your results

High A1C

A high A1C means your blood sugar has averaged too high over the last few months. At 6.5% or above, it points to diabetes; between 5.7% and 6.4%, to prediabetes — a stage where the trajectory is still very much reversible. The longer A1C stays elevated, the greater the risk of damage to the eyes, kidneys, nerves, heart, and blood vessels.

"My A1C is high but my fasting glucose is normal." Common, and not a contradiction. Your fasting glucose can be fine while post-meal spikes push your daily average up. A1C integrates the whole 24 hours, catching excursions a single fasting draw never sees. The reverse also happens: a normal A1C can sit on top of wide daily swings. That's why clinicians read A1C alongside glucose and, when insulin resistance is the question, insulin.

Low A1C

A low A1C in someone with diabetes usually means good control — but not always. It can also reflect frequent hypoglycemia averaging out the highs, which is why ADA guidance pairs glycemic goals with hypoglycemia assessment rather than chasing the lowest number.³ And an unexpectedly low result raises the next question, below: is it even accurate?

What falsely raises or lowers your A1C

A1C depends on two things: your blood sugar and how long your red blood cells live. Anything that shortens red cell survival gives glucose less time to attach, dragging A1C down; anything that makes red cells older than average pushes it up. Neither has anything to do with your diabetes control.⁷

Falsely lower A1C: hemolytic anemia and other causes of shortened red cell survival; recovery from acute blood loss and recent transfusion; erythropoietin (EPO) or newly started iron supplementation, which flood the blood with young red cells; advanced kidney disease and cirrhosis; pregnancy; high altitude; sickle cell disease.⁷⁸

Falsely raise A1C: iron-deficiency anemia — NGSP notes it "is associated with higher HbA1c"; thalassemia; vitamin B12 deficiency; splenectomy; and some medications, including certain immunosuppressants and protease inhibitors.⁷⁸

The iron point deserves emphasis because it is so common. If your ferritin is low, your A1C may read higher than your true average glucose — and it can fall after iron replacement without your blood sugar changing at all. Anyone whose A1C is being read near the 5.7% or 6.5% line should have their hemoglobin, MCV, and iron status looked at in the same breath.

Hemoglobin variants: the nuance most articles get wrong

Since A1C is hemoglobin, it's fair to ask what happens when yours isn't the standard adult type. This matters in the U.S.: more than 2 million Americans live with sickle cell trait, and about 1 in 13 Black babies born in the United States have it.¹⁵ Hemoglobin C, D, and E traits also occur. Here the sources genuinely diverge:

• NIDDK warns that "some A1C tests give falsely high or low readings" in people with hemoglobin variants, more common in those of African, Mediterranean, or Southeast Asian descent.¹
• NGSP — the body that certifies A1C assays — publishes a method-by-method table showing that across the ~20 certified methods listed, HbC, HbE, and HbD traits show no clinically significant interference, and HbS (sickle cell) trait shows conflicting data for a single method. Elevated fetal hemoglobin (HbF) does interfere with several methods.¹⁶

The reconciliation: modern NGSP-certified assays have largely solved trait interference, so someone with sickle cell trait can usually trust an A1C run on an appropriate method — NGSP's advice is simply that "your lab should use a method that does not show interference." The real problem lies elsewhere. In the homozygous and compound disease states — HbSS, HbCC, HbSC — NGSP is unambiguous that results "must be interpreted with caution given the pathological processes, including anemia, increased red cell turnover, and transfusion requirements."⁷¹⁷ The danger is not the variant molecule confusing the machine; it is the shortened red cell lifespan and transfusions.

Even in unremarkable hemoglobin, population differences remain unresolved: a GRADE substudy found A1C ran 0.2 to 0.6 percentage points higher in non-Hispanic Black than non-Hispanic White participants at the same CGM-measured glucose.¹⁸ No U.S. guideline applies different cutoffs by ancestry — but knowing the debate is live is part of reading your result honestly.

When A1C can't be trusted, what replaces it? Fingerstick glucose logs, CGM and time in range, or fructosamine / glycated albumin, which report over ~2–3 weeks and don't depend on red cell lifespan. In advanced kidney disease especially, CGM is increasingly preferred because A1C misleads there.¹⁹²⁰

How to lower your A1C

Lowering A1C means improving your average glucose over time, and because it's an average over months, it moves slowly. Expect 2 to 3 months before a change shows up.

• Food — fewer refined carbohydrates and sugar-sweetened drinks, more fiber, vegetables, and whole grains, on a regular meal rhythm.
• Physical activity, which improves insulin sensitivity — the DPP target was at least 150 minutes per week.
• Weight loss if you carry extra weight — the DPP aimed for at least 7% of body weight.
• Medication, adjusted by your clinician. Several drug classes now benefit the heart and kidneys too.

The evidence in prediabetes is unusually strong: in the Diabetes Prevention Program, a structured lifestyle program cut the incidence of type 2 diabetes by 58% over ~3 years versus 31% for metformin — lifestyle beat the drug.²¹ Never change a diabetes medication on your own, and set your target with your clinician, not from a chart on the internet — including this one.

When to see a doctor

Talk to your primary care provider if your A1C is 5.7% or above, if it is rising across tests, or if you have symptoms of high blood sugar — excessive thirst, frequent urination, unexplained weight loss, blurred vision, slow-healing wounds. Mention anything that could distort the result: anemia, iron deficiency, sickle cell trait or disease, thalassemia, kidney disease, pregnancy, a recent transfusion, or EPO. If you're 35 to 70 with overweight or obesity, ask about screening even if you feel fine — nearly 1 in 4 U.S. adults with diabetes don't know they have it.⁵

Recent research

According to recent publications indexed on PubMed:

• The thresholds are stable, the targets aren't. The ADA's 2026 Standards of Care reframe glycemic goals around individualization and hypoglycemia risk rather than one universal target — a direction set by the 2022 ADA/EASD consensus — while the diagnostic cutoffs stay put.³¹²
• A1C is not perfectly race-neutral. In 1,454 U.S. patients, the GRADE CGM substudy found the relationship between average glucose and A1C differs across racial groups; glycated albumin showed the same pattern, arguing the effect is biological rather than an assay artifact.¹⁸ (Nathan DM et al., Diabetes Care, 2024 — DOI.)
• Beyond A1C: time in range. International consensus defines time in range targets from CGM as a complement to A1C — especially where A1C is unreliable, as in kidney disease.¹⁴²⁰¹⁹

These findings concern diagnosis and monitoring; they do not authorize self-medication and do not replace your physician's advice.

Get your A1C interpreted by AI DiagMe

An A1C is never read alone: its meaning depends on your glucose, your hemoglobin and ferritin, your kidney function, your medications, and your context — and, as this guide shows, several common conditions can make the number itself wrong. That cross-referencing is what gives the result its real value.

👉 AI DiagMe interprets your lab results — blood, urine, or stool — in plain language, taking your whole profile into account. An informational service that does not provide a diagnosis and complements, never replaces, your physician.

Frequently asked questions

Bottom line

A1C is the memory of your blood sugar over 2–3 months, measured without fasting. Remember the chart — below 5.7% normal, 5.7–6.4% prediabetes, 6.5% or above diabetes — the common but personalized target of under 7%, and the eAG translation (7% ≈ 154 mg/dL). Remember too that it is hemoglobin you're measuring: iron deficiency, anemia, hemolysis, kidney disease, pregnancy, and transfusion can all make the number lie, and sickle cell disease — far more than trait — puts it out of reach. If you land in the prediabetes band, the news is good: lifestyle change cut progression by 58% in the DPP. No value is read in isolation — what counts is all your markers together with your profile, exactly what AI DiagMe provides.

Sources

Official sources and peer-reviewed publications (PubMed) used for this guide: