A normal A1C is below 5.7%. Prediabetes runs from 5.7% to 6.4%, and 6.5% or higher is in the diabetes range. The bands are the same for women, men and every adult age. Many adults with diabetes aim for under 7%, about 154 mg/dL average glucose, while frail older adults may aim for 7% to 8%. Anemia, low iron and hemoglobin variants can skew the result.

Normal A1C levels are below 5.7% for adults who do not have diabetes. We read the A1C pages from the NIH, the CDC and MedlinePlus, the 2008 study that turned A1C into average glucose, the 2026 American Diabetes Association goals, the ISPAD guideline for young people, and NIH guidance on what can throw the test off. You can see how we research health topics.
The short answer: under 5.7% is normal, 5.7% to 6.4% is prediabetes, and 6.5% or higher is in the diabetes range. Many adults with diabetes aim for under 7%, which equals an average glucose of about 154 mg/dL.[1][4][5]
| What it measures | The share of hemoglobin with glucose attached, which reflects about the past 3 months[1] |
|---|---|
| Normal | Below 5.7%[1][3] |
| Prediabetes | 5.7% to 6.4%[1] |
| Diabetes range | 6.5% or above, usually confirmed with a repeat A1C or a glucose test[1][2] |
| Common goal with diabetes | Under 7% for many nonpregnant adults; 7% to 8% or higher for some[4][1] |
| Fasting | Not needed[1][3] |
| Can mislead | Anemia, very low iron, kidney or liver disease, blood loss, transfusion, hemoglobin variants and pregnancy[1][9] |

What are normal A1C levels?
A normal A1C level is below 5.7%. The NIH, the CDC and MedlinePlus all use the same three bands for adults: below 5.7% is normal, 5.7% to 6.4% is prediabetes, and 6.5% or above is in the diabetes range. The number is the percentage of your hemoglobin that has sugar stuck to it.[1][2][3]
| A1C result | What it means | Estimated average glucose |
|---|---|---|
| Below 5.7% | Normal | Under about 117 mg/dL |
| 5.7% to 6.4% | Prediabetes | About 117 to 137 mg/dL |
| 6.5% or above | Diabetes range | About 140 mg/dL or higher |
Ranges come from NIDDK and the CDC; the glucose column uses the ADAG formula explained in the next section.[1][3][5]
Inside the prediabetes band, risk climbs with the number. The higher the A1C, the greater the chance of developing diabetes.[1] A first result in the diabetes range is usually checked again, or confirmed with a fasting glucose test or an oral glucose tolerance test.[2]
For a diagnosis, the blood should come from a vein and go to a lab that uses an NGSP-certified method.[1] If your number came back high and you feel unwell, our guide to high blood sugar symptoms explains which signs matter.
How does A1C convert to average blood sugar?
Estimated average glucose, or eAG, turns your A1C percentage into mg/dL, the unit on a home meter. The formula comes from a 2008 study of 507 people: eAG equals 28.7 times A1C, minus 46.7. An A1C of 7% works out to about 154 mg/dL, and 8% to about 183 mg/dL.[5][4]
| A1C | Estimated average glucose (mg/dL) |
|---|---|
| 5.0% | 97 |
| 5.5% | 111 |
| 5.7% | 117 |
| 6.0% | 126 |
| 6.5% | 140 |
| 7.0% | 154 |
| 8.0% | 183 |
| 9.0% | 212 |
| 10.0% | 240 |
We worked out each row from the ADAG formula; the 6%, 7%, 8% and 9% rows match the ADA table.[5][4] In the ADAG study, each person gave about 2,700 glucose readings over 3 months from finger sticks and continuous monitors. The link held across age, sex, diabetes type, and race or ethnicity.[5]
Your eAG will not match any single meter reading, because it is a long-term average.[1] If you wear a continuous glucose monitor, its report may show a glucose management indicator (GMI) instead. GMI equals 3.31 plus 0.02392 times your mean sensor glucose, so a mean of 150 mg/dL gives a GMI of 6.9%.[6] The GMI authors note that people get confused when GMI and the lab A1C disagree. For a diagnosis, the lab A1C is the number that counts.[6][1]
What are normal A1C levels by age?
The cutoff for a normal A1C is the same at every adult age: below 5.7%. Still, A1C drifts up with age in people without diabetes. A 2008 analysis of 5,743 adults found it rose about 0.01 points a year, even in people whose glucose tolerance tests were normal.[1][7]
| Age group | Framingham Offspring Study | NHANES 2001 to 2004 |
|---|---|---|
| Under 40 | 6.0% | 5.6% |
| 70 and older | 6.6% | 6.2% |
The authors said more research is needed before anyone uses age-specific cutoffs. NIDDK and the CDC still publish one set of bands for all adults.[7][1][3]
Children and teens with diabetes
The 2022 ISPAD guideline recommends an A1C under 7.0% (53 mmol/mol) for all young people with diabetes. It allows a lower goal of 6.5% (48 mmol/mol) early in type 1 diabetes, or when a family has a continuous monitor, automated insulin delivery and a skilled care team.[8] ISPAD kept 7.0% as the main goal because many families around the world lack that kind of access. Finger-stick glucose targets for children and teens are set to match an A1C under 7%.[8]
Older adults
The National Institute on Aging notes that A1C may not be accurate in people with some other health conditions, so doctors may add other tests.[10] Goals are often looser in later life. NIDDK says an A1C between 7% and 8%, or even higher, may suit people with limited life expectancy, long-standing diabetes or severe low blood sugar.[1] The ADA also allows less strict goals for people who are frail. Your care plan should say how often you get the A1C test.[4][10]
Are normal A1C levels different for women and men?
No. The bands for normal, prediabetes and diabetes are the same for women and men. NIDDK, the CDC and MedlinePlus publish one set of numbers for all adults. Pregnancy is the main exception for women, because pregnancy itself changes A1C and doctors rely on glucose readings instead.[1][2][3][9]
The ADAG study also found that the link between A1C and average glucose did not differ by sex.[5] So a woman over 50 or over 60 reads her result against the same 5.7% and 6.5% lines as a man of the same age.
For pregnancy, NIDDK says most people with diabetes should aim for an A1C of 6.5% or lower just before they conceive. Once pregnant, A1C is not the best way to track glucose.[9] The glucose targets health care professionals most often use during pregnancy are:[9]
- Fasting, before meals, at bedtime and overnight: 70 to 95 mg/dL
- 1 hour after eating: 110 to 140 mg/dL
- 2 hours after eating: 100 to 120 mg/dL
- Continuous monitor range: 63 to 140 mg/dL
An A1C may be checked early in pregnancy to look for diabetes that started before it. Gestational diabetes is screened with glucose tests, usually between 24 and 28 weeks.[1]
What A1C level is recommended for people with diabetes?
For many nonpregnant adults with diabetes, the ADA goal is an A1C under 7%. Some people aim for under 6.5% if they can get there without low blood sugar. Others, such as people with frequent lows, advanced complications or limited life expectancy, may be safer between 7% and 8% or a little higher.[4][1]
| Group | Usual A1C goal |
|---|---|
| Most nonpregnant adults with diabetes | Under 7% (53 mmol/mol)[4] |
| Adults who can reach it without low blood sugar | Under 6.5%[4][3] |
| Limited life expectancy, long-standing diabetes, severe lows or advanced complications | 7% to 8%, or higher[1] |
| Children, teens and young adults with diabetes | Under 7.0%; 6.5% in some settings[8] |
| Planning a pregnancy with diabetes | 6.5% or lower just before pregnancy[9] |
| Adults without diabetes | Below 5.7% is normal; this is a range, not a treatment goal[1] |
Some people with diabetes can lower their risk of complications by keeping A1C under 7%. Pushing lower is unsafe if it brings on hypoglycemia.[1] So you and your care team set the goal together, based on your health.
People with diabetes should have an A1C test at least twice a year. The ADA suggests about every 3 months when you are not at goal or after a treatment change.[1][4] If you are working on your number through diet, our low glycemic foods list shows swaps with published glycemic index values.
Can an A1C test be wrong?
Yes. Anything that changes how long red blood cells live, or the type of hemoglobin you carry, can push A1C falsely up or down. It is also imprecise: NIDDK says a 6.8% result could read 6.4% to 7.2% on a repeat of the same sample.[1]
| Factor | Effect on the result |
|---|---|
| Very low iron, such as iron-deficiency anemia | Can be falsely high; studies are mixed[1][12] |
| Recent blood loss, transfusion, hemodialysis or erythropoietin | Can change A1C[1] |
| Sickle cell disease and other hemoglobin variants | Falsely high or low with some test methods[1][11] |
| Kidney failure or liver disease | Can give false results[1][2] |
| Pregnancy | Changes A1C; glucose tests are preferred[9] |
People of sub-Saharan African, Mediterranean, South or Southeast Asian descent more often carry variants such as HbS or HbE. About 8% of African American babies are born with sickle cell trait.[11] Reliable methods exist for most variants, but A1C should not be used at all with HbSS, HbCC or HbSC.[11]
A 2026 review of 25 reports found the effect of low iron on A1C inconsistent in size and direction, though it may push people near a cutoff into the wrong band.[12]
A result is worth questioning when it does not match your meter or monitor, when it is below 4% or above 15%, or when it jumps after a lab switches methods.[11] Your doctor can then order a different A1C method or a glucose test.[1]
What research are the A1C numbers based on?
The A1C bands and conversions rest on a handful of large studies. ADAG tied A1C to thousands of measured glucose values. Later work tracked how A1C rises with age and what very low values might mean. The table lists who was studied and what each study found.[5][7][13]
| Study | Who was studied | What it found |
|---|---|---|
| ADAG, 2008[5] | 507 people at 10 centers, followed for 3 months | Average glucose equals 28.7 times A1C, minus 46.7; same across age, sex and race |
| Pani, 2008[7] | 5,743 adults without diabetes in two US cohorts | A1C rose 0.010 to 0.014 points per year of age |
| Bergenstal, 2018[6] | Data from recent continuous monitor trials | GMI equals 3.31 plus 0.02392 times mean glucose |
| Carson, 2010[13] | 14,099 US adults without diabetes, median 8.8 years | A1C under 4.0% was linked to higher death risk than 5.0% to 5.4% |
| Iron review, 2026[12] | 25 reports in adults without diabetes | Effect of iron deficiency on A1C was inconsistent |
The Carson study asked whether A1C can be too low. Its risk ratio was 3.73, and it fell to 2.90 after the authors adjusted for other factors.[13] People with very low A1C also had higher red cell volume, ferritin and liver enzymes. This was an observational study, so it cannot show that a low A1C causes harm.[13]
What should you do after your A1C result?
Match your next step to your band, then ask your clinician what goal fits you. A prediabetes result usually means a retest every 1 to 2 years, while a first result in the diabetes range needs a confirming test. If you already have diabetes and sit above your goal, expect a recheck in about 3 months.[2][4]
| If your A1C is | Usual next step |
|---|---|
| Below 5.7% and you have no diabetes | Normal; ask how often to retest, since the CDC advises testing from age 45[1][2] |
| 5.7% to 6.4% | Prediabetes; retest every 1 to 2 years[2][3] |
| 6.5% or above for the first time | Repeat A1C or a fasting or glucose tolerance test[2] |
| Above your goal with diabetes | Recheck about every 3 months and review treatment[4] |
| Out of line with your meter or monitor | Ask about anemia, iron, variants or another test method[1][11] |
When to call your doctor
Book a visit, or call your care team, if any of these apply to you.
- Your result is 6.5% or higher and you have not been diagnosed with diabetes[1]
- You are pregnant or planning a pregnancy and have diabetes[9]
- You have frequent low blood sugar while aiming for a tight goal[1]
- Your A1C and your daily readings tell different stories[11]
- You have prediabetes and have not been retested in 2 years[2]
- Your monitor’s GMI and your lab A1C are far apart[6]
What doctors use first
To diagnose diabetes, doctors use A1C alongside a fasting plasma glucose test or an oral glucose tolerance test, which need at least 8 hours of fasting. With clear symptoms, a random glucose test may be used.[1] Providers often use more than one test before they confirm a diagnosis.[2] A supplement plays no role in diagnosis or in setting an A1C goal. If you are curious about one, our article on cinnamon for blood sugar covers what the trials found and where they fall short.
Checklist before your next A1C test
- You can eat as usual unless other tests, such as cholesterol, need fasting[3]
- Bring your last A1C so you can compare the trend
- Tell the clinician about anemia, low iron, kidney or liver disease, recent blood loss or a transfusion[1]
- List your medicines, including any opioids or HIV drugs[2]
- If you wear a monitor, bring your mean glucose and GMI[6]
- Ask what A1C goal fits your age and health[2]
- Ask whether the lab will report your eAG next to the A1C, as some labs do[1]
Bottom line
A normal A1C is below 5.7%, prediabetes runs from 5.7% to 6.4%, and 6.5% or higher is in the diabetes range. These lines are the same for women and men and for every adult age, even though A1C creeps up slowly as people get older.[1][7]
For most people with diabetes, under 7% is a sensible starting goal. For frail older adults we think a looser target is often the wiser choice, because a goal that brings on lows can be unsafe.[4][1] If a result surprises you, ask for a repeat or a glucose test before acting on it. More guides sit in our prediabetes and diabetes section.
Questions people also ask
Does an A1C test require fasting?
No. You can eat before an A1C test, and blood can be drawn at any time of day. Your doctor may order other tests at the same visit, such as cholesterol or fasting glucose, and those may need you to fast first.
Can A1C be tested at home?
Home meters and monitors measure your glucose at one moment, which is different from an A1C. For a diagnosis, NIDDK says the blood should come from a vein and go to a lab that uses an NGSP-certified method.
Can A1C change in a month?
Partly. A1C reflects about 3 months because red blood cells live that long, but glucose from the past 30 days has the biggest effect. That is why people working on a goal usually retest after about 3 months.
Is an A1C of 5.7 bad?
An A1C of 5.7% is the lowest number in the prediabetes range. It means a higher risk of type 2 diabetes, not a diagnosis. Most people at this level are retested every 1 to 2 years and talk with their doctor about next steps.
Can A1C be too low?
In people without diabetes, a large US study linked an A1C under 4.0% with higher death risk, but it could not prove cause. In diabetes, chasing a very low number can bring dangerous lows, so goals are set to stay safe.
Sources
- The A1C Test & Diabetes, National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK), NIH
- Hemoglobin A1C (HbA1c) Test, MedlinePlus, National Library of Medicine
- A1C Test for Diabetes and Prediabetes, Centers for Disease Control and Prevention (CDC)
- 6. Glycemic Goals, Hypoglycemia, and Hyperglycemic Crises: Standards of Care in Diabetes 2026, American Diabetes Association, Diabetes Care, 2026
- Translating the A1C assay into estimated average glucose values, Nathan DM et al., Diabetes Care, 2008
- Glucose Management Indicator (GMI): A New Term for Estimating A1C From Continuous Glucose Monitoring, Bergenstal RM et al., Diabetes Care, 2018
- Effect of aging on A1C levels in individuals without diabetes, Pani LN et al., Diabetes Care, 2008
- ISPAD Clinical Practice Consensus Guidelines 2022: Glycemic targets and glucose monitoring for children, adolescents, and young people with diabetes, de Bock M et al., Pediatric Diabetes, 2022
- Pregnancy if You Have Diabetes, NIDDK, NIH
- Diabetes in Older People, National Institute on Aging (NIA), NIH
- Sickle Cell Trait & Other Hemoglobinopathies & Diabetes, NIDDK, NIH
- Iron Deficiency, Iron Deficiency Anemia, and HbA1c Levels in Nondiabetic Adults: A Systematic Review and Meta-Analysis, Cureus, 2026
- Low hemoglobin A1c and risk of all-cause mortality among US adults without diabetes, Carson AP et al., Circulation: Cardiovascular Quality and Outcomes, 2010
This article is general information, not medical advice. These statements have not been evaluated by the FDA, and supplements are not intended to diagnose, treat, cure or prevent any disease. Talk to your doctor or pharmacist before you start a supplement. How we research.


