A1c to Average Glucose Calculator

Glucose in the bloodstream attaches irreversibly to haemoglobin inside red blood cells. Because red cells survive about three months, the proportion of glycated haemoglobin at any moment is a weighted record of glucose exposure over that window — with the most recent month contributing roughly half the result.

The ADAG study measured that relationship directly rather than assuming it. Its regression, eAG = 28.7 × A1c − 46.7, is close to linear across the clinically useful range: each 1% of A1c corresponds to about 28.7 mg/dL, or 1.6 mmol/L, of average glucose.

The American Diabetes Association uses three diagnostic bands, applied to an A1c measured by an NGSP-certified assay standardised to the DCCT reference.

A diagnosis is not made on one result. The ADA asks for two abnormal tests — either two separate samples, or two different tests on the same sample, such as A1c plus a fasting glucose. Symptoms with a random glucose of 200 mg/dL or more are the exception.

Prediabetes is not a milder diabetes; it is a risk state. Roughly 5 to 10% of people with prediabetes progress each year, and structured lifestyle programmes cut that risk substantially — which is why the 5.7% threshold exists at all.

The ADA gives below 7% as a general target for many non-pregnant adults with diabetes, on the evidence that it reduces microvascular complications. That is a starting point for a conversation, not a rule.

• A tighter target, such as below 6.5%, may suit someone with a short diabetes duration, long life expectancy, and no significant cardiovascular disease — if it can be reached without hypoglycaemia.
• A looser target, such as below 8%, is often more appropriate with a history of severe hypoglycaemia, advanced complications, limited life expectancy, or extensive comorbidity.
• Pregnancy targets are different and considerably tighter, and are set by the obstetric team.
• An A1c above 9% is widely used as a quality measure for poor glycaemic control and usually prompts a review of the whole treatment plan.

A1c also says nothing about variability. Two people at 7.0% can have very different days — one steady around 154 mg/dL, the other swinging between hypoglycaemia and 300. Time in range from a continuous glucose monitor captures what A1c cannot.

The test assumes red blood cells live a normal lifespan. Anything that shortens or lengthens that skews the result independently of glucose:

• Shortened red cell survival — haemolytic anaemia, recent transfusion, sickle cell disease, splenomegaly, advanced kidney disease, erythropoietin treatment — reads falsely low.
• Iron deficiency, vitamin B12 deficiency, and asplenia lengthen red cell survival and read falsely high.
• Some haemoglobin variants interfere with particular assays. Laboratories can switch to a variant-tolerant method when this is suspected.
• Pregnancy shifts A1c downward from the second trimester, which is one reason it is not used for gestational diabetes diagnosis.

Where A1c is unreliable, fructosamine, glycated albumin, or direct glucose monitoring are used instead. A result that clearly contradicts a patient's own meter readings is worth questioning rather than accepting.

There are two scales. NGSP reports A1c as a percentage of total haemoglobin and is standard in the US. IFCC reports it in mmol per mol of haemoglobin and is standard in the UK, most of Europe, and Australia. They measure the same thing on different axes.

The master equation linking them is IFCC = (NGSP% − 2.15) × 10.929. A familiar landmark: 6.5% is 48 mmol/mol, and 7.0% is 53 mmol/mol. The IFCC numbers are larger and have no decimal, which is partly why the scale was adopted — it removes the confusion between an A1c percentage and a glucose percentage.