HbA1c Explained: Diabetes, Prediabetes, and AI Interpretation

What HbA1c measures

HbA1c — short for hemoglobin A1c, or glycated hemoglobin — is the percentage of your hemoglobin molecules that have glucose permanently attached to them. Hemoglobin is the iron-rich protein packed inside red blood cells that carries oxygen from your lungs to every tissue in your body. When glucose circulates in your bloodstream, a small fraction of it binds spontaneously and irreversibly to hemoglobin in a process chemists call non-enzymatic glycation. The higher your average blood glucose over time, the more of your hemoglobin becomes glycated, and the higher your HbA1c climbs.

The reason this single number is so clinically powerful comes down to red-cell biology. A red blood cell lives for roughly 90 to 120 days before the spleen recycles it. Because glycation is permanent for the life of the cell, the pool of red cells circulating at any moment carries a chemical memory of the glucose they have been bathed in. Your HbA1c therefore reflects your average blood glucose over the preceding two to three months, weighted toward the most recent four to six weeks (roughly half the signal comes from the last month, because newer cells outnumber the oldest ones). Unlike a fasting glucose reading, which is a single snapshot that can be thrown off by last night's dinner, a stressful commute, or a poor night's sleep, HbA1c smooths out day-to-day noise into one stable, integrated figure.

HbA1c is reported in one of two unit systems. In the United States and much of the world it is expressed as a percentage (for example, 5.5 %), a convention that dates to the Diabetes Control and Complications Trial (DCCT). Many countries, following the International Federation of Clinical Chemistry (IFCC), now also report it in millimoles per mole (mmol/mol). The two scales measure the same thing; we explain how to convert between them below. Throughout this guide we use the DCCT percentage because it remains the most widely recognized format on lab reports and in the diagnostic criteria published by the American Diabetes Association (ADA) and the World Health Organization (WHO).

Why clinicians rely on HbA1c

Before HbA1c became a diagnostic test, doctors relied almost entirely on fasting glucose and the oral glucose tolerance test (OGTT). Both remain useful, but both are logistically awkward. Fasting glucose requires the patient to arrive having eaten nothing for eight hours, which is inconvenient and frequently violated in the real world. The OGTT requires drinking a sugary solution and sitting in a lab for two hours with repeated blood draws. HbA1c requires neither: it can be drawn at any time of day, in the fed or fasted state, and a modest deviation in the hours before the test barely moves the result.

In 2009 an International Expert Committee, and shortly afterward the ADA and WHO, formally endorsed HbA1c ≥ 6.5 % as a diagnostic criterion for diabetes. That decision reflected a decade of evidence that HbA1c predicts the microvascular complications of diabetes — retinopathy (eye damage), nephropathy (kidney damage), and neuropathy (nerve damage) — at least as well as glucose-based tests, and often more reproducibly. The landmark DCCT (type 1 diabetes) and UK Prospective Diabetes Study (UKPDS, type 2 diabetes) both demonstrated that lowering HbA1c reduces the long-term risk of these complications, cementing it as the primary target for glycemic management worldwide.

For people already living with diabetes, HbA1c is the yardstick of long-term control. It is the number that endocrinologists, primary care physicians, and diabetes educators track over years to judge whether a treatment plan is working. It is also the outcome that virtually every major diabetes drug trial is measured against. In short, HbA1c is simultaneously a screening test, a diagnostic test, and a monitoring test — a rare trifecta in laboratory medicine that explains its central place in metabolic care.

How HbA1c is measured and standardized

HbA1c is measured on a blood sample using one of several analytical methods, most commonly high-performance liquid chromatography (HPLC), immunoassay, boronate affinity chromatography, or enzymatic assays. These methods physically or chemically separate glycated hemoglobin from non-glycated hemoglobin and report the ratio. The point-of-care devices found in some clinics and pharmacies can produce a result in minutes from a finger-prick sample, though laboratory venous testing remains the reference standard for diagnosis.

A critical piece of history explains why you can trust a result from one lab to mean the same thing at another. In the 1990s, HbA1c results were notoriously inconsistent between laboratories. The National Glycohemoglobin Standardization Program (NGSP) in the United States, working alongside the IFCC, established reference methods and a certification scheme that aligned assays to the DCCT and UKPDS trials. Today, laboratories using NGSP-certified methods produce results that are comparable within a tight margin. This standardization is precisely why a fixed diagnostic cutoff of 6.5 % is defensible internationally — something that is not true for many other lab tests, where reference ranges vary by assay and population.

For diagnosis, the ADA and WHO both stress that the test must be run in an accredited laboratory using an NGSP-certified, IFCC-standardized method — not a point-of-care device, which is acceptable for monitoring but not for making a first diagnosis. They also emphasize that, in the absence of unequivocal symptoms of high blood sugar, an abnormal result should be confirmed with a repeat test before a diagnosis of diabetes is made. A single elevated HbA1c is a reason to retest, not a verdict.

Diagnostic thresholds (2026)

The diagnostic cutoffs below reflect the criteria maintained by the American Diabetes Association in its annual Standards of Care and the World Health Organization, and they are unchanged in substance for 2026. They apply to non-pregnant adults using a standardized laboratory assay. Note that the same numbers that diagnose diabetes are different from the numbers used as treatment targets once you already have the condition — a distinction that confuses many patients.

The prediabetes band (5.7–6.4 %) is arguably the most important part of this table, because it is the window in which the trajectory toward type 2 diabetes can still be bent or reversed. The ADA uses 5.7 % as the lower bound; some other bodies, including the WHO, have historically been more conservative and reserve the term 'intermediate hyperglycemia' for a narrower range, while the UK's NICE guidance flags 6.0–6.4 % (42–47 mmol/mol) as 'high risk.' These differences matter less than the underlying principle: a result in the high-5s is a signal to act, not to panic.

The treatment targets are deliberately individualized. The ADA and the American Association of Clinical Endocrinology both stress that a target of < 7.0 % suits most non-pregnant adults, but that a stricter goal of < 6.5 % is reasonable for younger patients early in the disease who can reach it without dangerous low blood sugar (hypoglycemia), while a looser goal of < 8.0 % is safer for older adults, those with a history of severe hypoglycemia, advanced complications, or limited life expectancy. The 2008 ACCORD trial famously showed that pushing HbA1c too aggressively toward normal in high-risk type 2 patients increased mortality — a permanent reminder that lower is not always better.

Translating HbA1c into average glucose

Because HbA1c is a percentage and home glucose meters read in mg/dL (or mmol/L), the ADA popularized the concept of estimated average glucose (eAG) to put both numbers on the same scale. The relationship comes from the ADAG study, which correlated HbA1c with continuous glucose monitoring in hundreds of people. The formula is eAG (mg/dL) = 28.7 × HbA1c% − 46.7. This lets you see roughly what average blood sugar a given HbA1c corresponds to.

A vital caveat: eAG is an average, and averages hide variability. Two people can share an HbA1c of 7.0 % while one has steady glucose all day and the other swings between damaging highs and dangerous lows that cancel out on paper. This is why continuous glucose monitoring (CGM) metrics such as time-in-range increasingly complement HbA1c rather than replace it. HbA1c tells you the height of the tide; it says nothing about how rough the sea was. When you review your fasting glucose alongside HbA1c, you begin to see both the level and some of the shape of your glucose curve.

Causes of a high HbA1c

A genuinely elevated HbA1c means your average blood glucose has been running high. The overwhelming majority of the time this reflects insulin resistance and type 2 diabetes or its precursor state, but the differential is broader than many people realize. Understanding why the number is high is what turns a lab value into a plan.

Common and expected causes

• Type 2 diabetes and prediabetes — by far the most common reason, driven by insulin resistance from excess visceral fat, physical inactivity, and genetic predisposition.
• Type 1 diabetes — autoimmune destruction of insulin-producing beta cells, more common in younger patients but possible at any age (including latent autoimmune diabetes in adults, LADA).
• Steroid and other medications — glucocorticoids (prednisone), some antipsychotics, thiazide diuretics, and certain immunosuppressants raise glucose.
• Cushing's syndrome and acromegaly — endocrine disorders that oppose insulin.
• Chronic pancreatitis or pancreatic surgery — loss of insulin-producing tissue.
• Pregnancy-related hyperglycemia — though gestational diabetes is diagnosed with glucose testing, not HbA1c.

Falsely high HbA1c (glucose is actually normal)

Any condition that lengthens the average age of your red blood cells gives glucose more time to accumulate on hemoglobin and pushes HbA1c up even when your true glucose is fine. The classic example is iron-deficiency anemia: when iron is scarce, the bone marrow produces fewer new red cells, so the circulating population skews older and more glycated. Studies show iron deficiency can inflate HbA1c by several tenths of a percent, occasionally enough to cross a diagnostic line. This is one of the most important and under-recognized pitfalls in HbA1c interpretation, and it is why our analyzer cross-checks iron markers — you can read more in our guide to iron-deficiency anemia and review ferritin, the best single marker of iron stores.

• Iron, vitamin B12, or folate deficiency anemia — older red-cell population, falsely raising HbA1c.
• Asplenia (no spleen) — red cells survive longer without splenic clearance.
• Severe hypertriglyceridemia, hyperbilirubinemia, or uremia — can interfere with certain assay methods.
• Chronic alcohol use and lead poisoning — can spuriously elevate some measurements.
• Certain hemoglobin variants — depending on the assay, some variants read falsely high (others falsely low).

Causes of a low or falsely low HbA1c

A low HbA1c — say, below 5.0 % — is usually reassuring and simply reflects excellent glucose control. But an unexpectedly low result, especially one that clashes with high home glucose readings, is a red flag that the test is being fooled. The mechanism is the mirror image of the false-high scenario: anything that shortens red-cell lifespan or increases red-cell turnover gives glucose less time to attach, dragging HbA1c below the patient's true glucose level.

• Hemolytic anemia — accelerated red-cell destruction produces a young, minimally glycated red-cell population and a falsely low HbA1c.
• Recent significant blood loss or blood transfusion — dilutes glycated hemoglobin with fresh or donor cells.
• Recovery from anemia / erythropoietin (EPO) therapy / iron or B12 repletion — a burst of new red-cell production lowers the average cell age.
• Advanced chronic kidney disease — shortened red-cell survival and EPO treatment both pull HbA1c down (glucose-based tests or fructosamine may be preferred).
• Chronic liver disease and splenomegaly — increased red-cell turnover and sequestration.
• Pregnancy (second and third trimesters) — increased red-cell turnover lowers HbA1c, one reason it is unreliable for diagnosing gestational diabetes.
• Some hemoglobin variants — can read falsely low on affected assays.

Who is at risk and who should be screened

Type 2 diabetes and prediabetes are largely silent for years — the CDC estimates that a large share of people with prediabetes do not know they have it. That is exactly why screening asymptomatic adults matters. The US Preventive Services Task Force (USPSTF) recommends screening for prediabetes and type 2 diabetes in adults aged 35 to 70 years who are overweight or obese, and repeating it every three years if normal. The ADA casts a slightly wider net, recommending screening from age 35 in all adults, and earlier in anyone with additional risk factors.

Risk factors that warrant earlier or more frequent testing

• Overweight or obesity, especially abdominal (visceral) fat.
• A first-degree relative (parent or sibling) with diabetes.
• Physical inactivity and a diet high in refined carbohydrate and sugar-sweetened beverages.
• High-risk race or ethnicity (including South Asian, East Asian, Hispanic/Latino, African, and Indigenous populations, who develop insulin resistance at lower body weights).
• A history of gestational diabetes or delivering a baby over 9 lb / 4 kg.
• Polycystic ovary syndrome (PCOS).
• Hypertension, low HDL cholesterol, or high triglycerides — the components of metabolic syndrome.
• Known cardiovascular disease, non-alcoholic fatty liver disease, or use of glucose-raising medications such as chronic steroids.

Ethnicity deserves special emphasis. Major bodies including NICE and the ADA recognize that people of South Asian and several other ancestries develop type 2 diabetes at a lower body-mass index than white European populations, which is why some guidelines lower the BMI screening threshold for these groups. If you have a strong family history or belong to a higher-risk group, an HbA1c in the high-5s deserves more attention than the same number in a low-risk person — context that a good analyzer weighs automatically.

Where HbA1c misleads

No single blood test is perfect, and HbA1c has well-catalogued blind spots. The original version of this guide highlighted several, and they remain true; here we expand on each and add the mechanism so you can reason about your own result.

Two limitations are worth underlining. First, in hemoglobinopathies — inherited variants such as sickle-cell trait (HbS), HbC, and HbE, which are common in people of African, Mediterranean, Middle Eastern, and South-East Asian ancestry — the reliability of HbA1c depends entirely on which assay the lab uses. The NGSP publishes tables of which methods are affected by which variants. A person with an unexpected result and relevant ancestry should ask whether their lab's method is variant-sensitive.

Second, in pregnancy, HbA1c is not used to diagnose gestational diabetes; the OGTT is the standard, because red-cell turnover rises and shifts HbA1c downward, and because the timescale of pregnancy-related glucose changes is faster than HbA1c can track. In these and other situations, clinicians turn to fructosamine (which reflects glucose over the past two to three weeks and is independent of red-cell lifespan) or to direct glucose testing and CGM.

Related conditions and companion tests

HbA1c almost never travels alone. Because high blood sugar is one facet of a broader metabolic picture, its interpretation improves dramatically when read next to a handful of companion markers. This is the core of good metabolic medicine and the logic our analyzer follows.

• Fasting glucose — the classic partner test. When HbA1c and fasting glucose agree, confidence is high; when they disagree, one of them is likely being distorted (often HbA1c by a red-cell condition).
• Lipid panel — insulin resistance characteristically raises triglycerides and lowers HDL. A high HbA1c with high triglycerides and low HDL is the metabolic-syndrome signature. Our lipid panel guide explains how these move together.
• Complete blood count and iron studies — essential for detecting the anemias that distort HbA1c. See our CBC explainer.
• Kidney function (eGFR, creatinine, urine albumin) — diabetes is a leading cause of kidney disease, so a high HbA1c prompts a look at the kidneys.
• hs-CRP — chronic inflammation accompanies insulin resistance and adds cardiovascular risk; review C-reactive protein.
• Thyroid function — thyroid disorders influence metabolism and lipids; an abnormal TSH can complicate the metabolic picture (see our thyroid panel guide).

Conditions strongly associated with a high HbA1c include metabolic syndrome, non-alcoholic fatty liver disease (NAFLD/MASLD), polycystic ovary syndrome, obstructive sleep apnea, and cardiovascular disease. The AHA/ACC and the European Society of Cardiology (ESC) both treat diabetes and prediabetes as major, independent cardiovascular risk factors — which is why an elevated HbA1c should always trigger a cardiovascular risk conversation, not just a glucose one.

Personalized targets and how often to test

Once diabetes is diagnosed, HbA1c shifts from a diagnostic role to a monitoring one, and the question becomes: what is your target, and how often should you check? The answer is genuinely individual. The Endocrine Society, ADA, and NICE all endorse individualized targets that weigh the benefit of tight control (fewer long-term complications) against its main hazard (hypoglycemia).

Because HbA1c reflects roughly three months of glucose, testing it more often than every three months rarely adds information — the biology simply cannot change faster than that. If you have just started a new medication or overhauled your diet, waiting the full three months gives the number time to reflect the change. Checking it every few weeks and reacting to small moves is a common source of unnecessary anxiety.

Evidence-based ways to lower HbA1c

If your HbA1c sits in the prediabetes range, the evidence that you can change its course is strong and specific. The landmark Diabetes Prevention Program (DPP), funded by the NIH, showed that a structured lifestyle intervention — modest weight loss of about 7 % of body weight and 150 minutes of moderate activity per week — reduced progression from prediabetes to type 2 diabetes by 58 %, outperforming the drug metformin (31 %) in the overall population. That is one of the most important findings in preventive medicine, and it means the number in the high-5s is often reversible.

What actually moves HbA1c

• Weight loss, especially of visceral fat — the single most effective lever for insulin resistance. Even 5–7 % of body weight makes a measurable difference.
• Regular physical activity — both aerobic exercise and resistance training improve insulin sensitivity; muscle is a major glucose sink. The DPP target of 150 minutes per week is a reasonable floor.
• Dietary pattern — reducing refined carbohydrates, sugar-sweetened beverages, and ultra-processed foods, and emphasizing fiber, legumes, vegetables, and whole grains. Mediterranean-style and lower-carbohydrate patterns both have supporting evidence.
• Sleep and stress — chronic sleep deprivation and psychological stress raise cortisol and worsen glucose control; treating sleep apnea can improve HbA1c.
• Reducing alcohol and stopping smoking — smoking independently worsens insulin resistance and cardiovascular risk.

In real-world terms, sustained lifestyle change can lower HbA1c by roughly 0.5 to 1.5 percentage points over six months in many adults, with the largest gains in early prediabetes and early type 2 diabetes. When lifestyle alone is not enough, clinicians add medications — metformin remains first-line for type 2 diabetes, and newer classes such as GLP-1 receptor agonists and SGLT2 inhibitors offer additional cardiovascular and kidney benefits that the AHA/ACC and ESC now weigh heavily. Any medication decision belongs with your clinician.

How AI interprets HbA1c in context

A naive analyzer flags a single HbA1c of 5.8 % as prediabetes and stops. That is technically correct and clinically shallow. The value of an AI blood test analyzer — like the Kantesti AI engine that powers blood-test.life — is that it reads HbA1c the way an experienced clinician does: as one thread in a much larger fabric.

When our analyzer sees an HbA1c result, it does not evaluate it in isolation. It cross-references your fasting glucose to check whether the two agree, scans your complete blood count and iron studies for the anemias that falsely raise or lower HbA1c, considers your lipid profile and hs-CRP for the metabolic-syndrome pattern, and — most importantly — looks at your trend over time rather than a single snapshot. A stable 5.8 % held for three years carries a very different meaning from a 5.8 % that climbed from 5.3 % in eighteen months.

What the AI adds that a lone number cannot

• Cross-validation — checking HbA1c against glucose and red-cell markers to catch false highs and lows before they mislead you.
• Trend detection — recognizing the difference between a stable value and one that is quietly rising, which changes urgency.
• Context weighting — giving the same number more or less significance based on age, family history, ethnicity, BMI, and companion markers.
• Pattern recognition — spotting the metabolic-syndrome cluster (high HbA1c + high triglycerides + low HDL + high hs-CRP) that no single value reveals.
• Plain-language explanation — translating the finding into what it means for you and what to ask your clinician.

You can read more about the methodology on our how it works page, and about the broader philosophy in our pillar guide to the AI blood test analyzer. The goal is never to replace your doctor — it is to help you arrive at the appointment understanding your numbers, with better questions.

When to see a clinician

An HbA1c result is a prompt for a conversation, and some results make that conversation urgent. Seek medical advice — promptly, not eventually — in the following situations.

• Any HbA1c ≥ 6.5 %, which needs confirmation and a formal diabetes evaluation.
• An HbA1c in the prediabetes range (5.7–6.4 %), to plan prevention and rule out other risks.
• A result that disagrees with your home glucose readings or CGM, which suggests a red-cell condition distorting the test.
• Classic symptoms of high blood sugar — excessive thirst, frequent urination, unexplained weight loss, blurred vision, recurrent infections, or slow-healing wounds — regardless of your number.
• Symptoms of low blood sugar if you are on glucose-lowering medication — shakiness, sweating, confusion — which may mean your target is too tight.
• A rising trend across several tests, even within the normal range.

If you experience symptoms of a diabetic emergency — severe dehydration, deep or rapid breathing, fruity-smelling breath, drowsiness, or confusion — treat it as urgent and seek emergency care immediately. These can signal diabetic ketoacidosis or a hyperosmolar state, both of which are life-threatening.

Common myths about HbA1c

A normal HbA1c means I definitely do not have diabetes

Not always. HbA1c can be falsely lowered by hemolysis, recent blood loss, kidney disease, and pregnancy. If your glucose readings are high but your HbA1c looks fine, the test may be masking a real problem. This is exactly why guidelines allow diagnosis by glucose testing when HbA1c is unreliable.

I fasted before my HbA1c, so it should be lower

Fasting does not change HbA1c meaningfully. Unlike glucose, HbA1c reflects months of average blood sugar, not the last meal. You do not need to fast for it, and skipping breakfast will not improve the number.

One high result means I have diabetes

A single elevated HbA1c is a reason to repeat the test, not a diagnosis. The ADA and WHO require confirmation — either a repeat HbA1c or a second test — before diagnosing diabetes, unless you have unequivocal symptoms plus a clearly diabetic glucose.

If my HbA1c is good, my glucose control is perfect

A good HbA1c is an average and can hide wide swings between highs and lows. Two people at 7.0 % can have very different day-to-day glucose stability. Time-in-range from a CGM captures what HbA1c cannot.

Prediabetes always becomes diabetes

It does not. The Diabetes Prevention Program proved that lifestyle change can cut progression by more than half. Prediabetes is a warning with a wide-open exit, not a sentence.

Key takeaways

• HbA1c is the percentage of glucose-bound hemoglobin and reflects your average blood glucose over roughly three months — no fasting required.
• 2026 thresholds (ADA/WHO): normal < 5.7 %, prediabetes 5.7–6.4 %, diabetes ≥ 6.5 % confirmed on two occasions; most adults with diabetes aim for < 7.0 %.
• The same number can be falsely high (iron deficiency, older red cells) or falsely low (hemolysis, blood loss, kidney disease, pregnancy) — always read it alongside glucose and a CBC.
• Prediabetes is often reversible: modest weight loss and 150 minutes of weekly activity cut progression by more than half in the NIH Diabetes Prevention Program.
• HbA1c is a screening and monitoring tool, not a diagnosis or a medical device — confirm abnormal results with a clinician, especially if you have symptoms.
• A good AI analyzer reads HbA1c in context — against glucose, iron markers, lipids, inflammation, and your trend — not as a lone number.

Ready to see your own HbA1c interpreted alongside the rest of your panel? Explore the full biomarker library or view pricing to run your labs through the analyzer.

Frequently asked questions