An HbA1c result can be wrong while the laboratory did everything correctly. In the guides we publish here, A1c is the marker where this happens most: the number reads one way and the person's physiology reads another.
The Verdict
Why the Test Is Vulnerable in the First Place
HbA1c reports the percentage of hemoglobin that has glucose attached to it. Two quantities decide that percentage: how much glucose has been available, and how long the average red blood cell has been alive to collect it. The test is designed to measure the first. It cannot separate it from the second.
Red blood cells live about 120 days when everything is normal. Push that lifespan up and the cells accumulate more glucose, so the percentage rises. Push it down, or replace old cells with new ones, and the percentage falls. Neither movement involves your blood sugar at any point. Our page on what the HbA1c test measures covers the normal case. This one covers the situations where the assumption breaks.
The Conditions That Distort A1c, and Which Way
Seven situations account for nearly every misleading A1c a reader is likely to meet. Each is common, and most are visible on a full blood count that was probably drawn at the same time.
| Condition | Direction | What to check or order instead |
|---|---|---|
| Iron deficiency, with or without anemia | Falsely high | Ferritin with a full blood count. Correcting iron typically drops A1c by a few tenths of a percent on its own |
| B12 or folate deficiency | Usually falsely high | B12 and folate. Both slow red cell production, so the surviving cells are older and more glycated. The evidence here is thinner than for iron |
| Recent blood loss, a transfusion, or hemolysis | Falsely low | A full blood count and reticulocyte count. New cells have had less exposure to glucose |
| Chronic kidney disease | Usually falsely low | Fructosamine or glycated albumin. Shortened red cell survival and erythropoietin treatment both pull the number down |
| Pregnancy, second and third trimester | Falsely low | An oral glucose tolerance test, which is what gestational diabetes is diagnosed on |
| Hemoglobin variants: HbS, HbC, HbE, HbD trait | Either direction, method-dependent | Confirm the method against the National Glycohemoglobin Standardization Program (NGSP) interference table, or switch to fructosamine |
| Recently started iron, B12 or erythropoietin | Falsely low | Wait three months. A burst of new red cells resets part of the sample |
Iron deficiency deserves the most attention because it is both the commonest cause and the one that moves people across a diagnostic line. The NGSP lists iron deficiency among the conditions that raise HbA1c without any change in glucose, and reported falls after iron repletion alone sit in the range of 0.2 to 0.5 percentage points. A woman with heavy periods and a ferritin of 12 ng/mL who reads 5.9% may sit at 5.5% three months after correcting her iron, having changed nothing about how she eats.
Chronic Kidney Disease Pulls in Two Directions
Advanced kidney disease usually makes HbA1c read low. Red cells survive less time in uremia, and erythropoietin treatment adds a wave of young cells that have collected almost no glucose. Both effects push the percentage down while glucose is unchanged.
A smaller effect works the other way. Urea produces carbamylated hemoglobin, which some older measurement methods could not distinguish from glycated hemoglobin, reading it as A1c. Modern high-performance liquid chromatography and immunoassay platforms largely separate the two, so the downward pull dominates in current practice.
For anyone with an estimated glomerular filtration rate, or eGFR, under 30 mL/min/1.73m², treat HbA1c as an underestimate and ask about glycated albumin. The gap can reach a full percentage point at dialysis-level kidney function, which is the difference between a result that looks controlled and one that does not.
Hemoglobin Variants and the Method Problem
Sickle cell trait, hemoglobin C trait, hemoglobin E and hemoglobin D each change the protein being measured. Whether that changes your result depends on the analytical method as much as on the variant.
Ion-exchange chromatography separates hemoglobin species by charge, and a variant can co-elute with the A1c peak or shift away from it. Some immunoassays bind the variant's altered terminal region differently. Others are unaffected. The consequence is that the same blood can produce a 6.8% on one platform and a 5.9% on another, and neither laboratory has made an error.
Two steps resolve it. Ask which method the laboratory runs and check it against the NGSP interference table. If a variant is present and the method is affected, move to fructosamine or a glucose-based test permanently rather than switching laboratories and hoping.
What to Order When A1c Cannot Be Trusted
Four alternatives cover almost every case, and they differ in what window they see and what can distort them.
| Test | Window it covers | What it does well | Where it falls down |
|---|---|---|---|
| Fructosamine | Previous 2 to 3 weeks | Does not involve hemoglobin at all, so red cell problems cannot distort it | Distorted by low albumin, liver disease and thyroid disease |
| Glycated albumin | Previous 2 to 3 weeks | More specific than fructosamine and unaffected by red cells | Not offered by every laboratory, and affected by protein loss |
| Oral glucose tolerance test | One two-hour window | The diagnostic standard when A1c is unreliable, and what pregnancy uses | Two hours in a clinic, and a fasting requirement |
| Continuous glucose monitor | Previous 14 days | Shows the whole curve rather than an average, and prices have fallen sharply | Sensor bias of its own, and not a diagnostic test |
Fructosamine is the usual first substitute because it is widely available, inexpensive, and immune to every red cell problem above. Its own weakness is albumin: nephrotic syndrome, liver disease and thyroid disease all move it. Glycated albumin is the more specific version of the same idea where a laboratory offers it.
An oral glucose tolerance test remains the answer when the question is diagnostic rather than monitoring. It is what pregnancy uses, for exactly that reason, and it measures glucose handling directly instead of inferring it from a protein.
Who This Does Not Apply To
Most people do not need any of this. Take a normal full blood count and ferritin, no pregnancy, intact kidney function and no known hemoglobin variant. In that situation your HbA1c means what it says, and the alternatives above add cost without adding information.
Four groups should read the list carefully. Menstruating women, anyone with a family history of a hemoglobin variant, and people with chronic kidney disease make up three. The fourth is anyone whose A1c disagrees with their meter, their sensor or their previous results. In those groups a single A1c is a hypothesis rather than an answer.
What would change our answer is a routinely reported, standardised glycated albumin. It answers the same question as HbA1c over a shorter window and does not depend on red cells at all. If laboratories reported it as widely as they report A1c, the interference list above would stop mattering. Until then, order a ferritin and a full blood count alongside any A1c you plan to act on, and take all three to your appointment.
Frequently Asked Questions
Can hemoglobin A1c be falsely elevated?
Yes, and iron deficiency is the most common cause. When iron is short, the bone marrow makes fewer new red cells, so the cells in your sample are older on average and have spent longer collecting glucose. The result reads high while your actual glucose has not moved. B12 and folate deficiency do the same thing through the same route. Correcting a documented iron deficiency commonly lowers HbA1c by 0.2% to 0.5% with no change in diet, which is enough to move someone across the 5.7% prediabetes line in either direction.
What conditions make an A1c read falsely low?
Anything that shortens red cell life or floods the blood with new cells. Recent blood loss, a transfusion, hemolytic anemia, advanced chronic kidney disease, the second and third trimesters of pregnancy, and starting iron, B12 or erythropoietin treatment all produce a lower number than your glucose deserves. The risk here is the mirror image of a false high: a person with real hyperglycemia gets a reassuring 5.5% and no follow-up.
Does anemia affect HbA1c results?
It affects them in both directions, depending on the type. Anemia caused by low production, such as iron, B12 or folate deficiency, raises HbA1c because red cells survive longer. Anemia caused by destruction or loss, such as hemolysis or bleeding, lowers it because cells are replaced faster. A single word on a report saying "anemia" does not tell you which way the A1c has moved, so read it alongside the mean corpuscular volume, or MCV, and the ferritin that identify the cause.
Do sickle cell trait and other hemoglobin variants change an A1c result?
They can, and whether they do depends entirely on which method your laboratory runs. HbS, HbC, HbE and HbD trait interfere with some ion-exchange and immunoassay methods while leaving others unaffected. The National Glycohemoglobin Standardization Program publishes a table of which assays are affected by which variant. Ask the laboratory which method it uses before repeating the test elsewhere, because a second draw sent to the same platform reproduces the same error.
How do you check whether an A1c result is accurate?
Order a marker that does not depend on red cells. Fructosamine and glycated albumin both reflect the previous two to three weeks through glycated blood proteins, so hemoglobin problems cannot touch them. A large disagreement between HbA1c and fructosamine points at the red cells rather than at your glucose. A fasting glucose and an oral glucose tolerance test answer the underlying question directly, and a 14-day continuous glucose monitor average gives a picture neither blood test can.
Is A1c different across ethnic groups?
Yes, and the difference is large enough to matter at the diagnostic line. At the same measured average glucose, HbA1c runs roughly 0.2 to 0.4 percentage points higher in Black adults than in white adults, a gap reported consistently in NHANES and other United States population data. The mechanism is not settled and appears to involve red cell turnover rather than glucose handling. The practical effect is that a Black adult can cross the 6.5% threshold at a mean glucose that leaves a white adult below it, so a borderline result deserves a fasting glucose alongside it.
Does vitamin C or a supplement change an A1c reading?
High-dose vitamin C and vitamin E have been reported to lower measured HbA1c on some assay platforms, through interference with the measurement rather than any effect on glucose. The effect is smaller and less consistent than the red cell causes above. If you take gram-level doses of either, mention it to whoever ordered the test. Biotin, which distorts many other assays, is not a problem for HbA1c, and our page on the TSH blood test covers where biotin genuinely does cause trouble.
Should you repeat an A1c that looks wrong?
Repeating the same test at the same laboratory reproduces the same interference, so it settles nothing. If the result conflicts with your symptoms, your meter readings or your last result, the useful next step is a different marker: a fasting glucose, an oral glucose tolerance test, or fructosamine. Bring the full blood count and ferritin as well, because the answer is often visible there.
Related
- HbA1c: what the percentage measures
- Ferritin, the marker that falls first
- MCV and what it says about anemia
- eGFR and kidney function
- At-home A1c tests
- Normal vs optimal ranges
- False high potassium — the same problem on a different marker, where the artefact happens in the tube rather than in the blood