Where HbA1c misleads, and what to read beside it
Two people hand over the same HbA1c of 5.7%. One spends almost the whole day inside a narrow band of glucose. The other climbs and crashes three times a day and never finds out.
Two people hand over the same HbA1c result of 5.7%. One of them spends almost the whole day inside a narrow band of glucose, the sugar in the blood. The other climbs and crashes three times a day and never finds out. Then a third person changes nothing about how she eats, takes iron tablets for six weeks, and watches her HbA1c fall half a point.
By the end of this piece you will know what HbA1c physically measures, the two quite different ways it can mislead, and what is worth reading beside it.
What is the number actually counting?
Red blood cells carry a protein called haemoglobin, which moves oxygen around your body. Glucose in the blood sticks to haemoglobin slowly, steadily and permanently. Once a sugar molecule attaches, it stays attached for as long as that red cell survives. HbA1c is simply the percentage of haemoglobin carrying that sugar mark.
The standard description then goes soft. Almost everyone repeats that HbA1c is a three-month average. It is not quite one. Red cells in your circulation are of every age at once, from a day old to around four months. The older ones are steadily being retired. The result is a weighted figure, not a flat one. Roughly half of an HbA1c is set by the most recent month, and only about a quarter by the oldest. That is why the value moves sooner than people expect, and why a good fortnight before a blood draw genuinely shows in it.
That weighting is the honest part of the test. The trouble starts with the other word: average.
Why do two people with the same average differ?
An average stands in for a whole day of readings and discards the shape completely. Two people can reach an identical average glucose by entirely different routes. One holds a narrow band all day. The other spikes high after every meal and drops low in the hours between. The peaks and troughs cancel out into a respectable-looking number.
The same HbA1c can sit over a calm profile or over a day of large swings, and those swings carry a risk the average cannot show. The peaks strain the lining of blood vessels. The lows between them are their own event. In someone on glucose-lowering medication, the lows are the more urgent one. Worse, lows drag the average downward. So the most turbulent glucose can produce the most reassuring number.
Averaging is the failure most people have already heard of. The second one catches out careful readers.
What could move the number without any glucose moving?
The answer is the red cell itself. The sugar mark builds up over a cell’s lifetime, so the mark depends on how long the cell has been circulating. Cells that live eighty days cannot carry as much attached sugar as cells that live a hundred and twenty, even at identical glucose. Anything that shortens red cell survival lowers HbA1c independently of glucose, and anything that lengthens it raises the value.
Several things shorten red cell survival. Recent blood loss does. So do chronic kidney disease, pregnancy, and conditions where red cells break down early, which doctors call haemolysis. Correcting an iron deficiency does it too. The new iron floods the circulation with young cells and can pull HbA1c down for months, while nothing about the person’s metabolism has improved. Other things lengthen survival. Untreated iron deficiency, and deficiencies of vitamin B12 or folate, all raise HbA1c without raising glucose.
The measurement method adds another layer. Some people carry inherited variants of haemoglobin, such as sickle trait. These variants interfere with certain laboratory methods, shifting the result in either direction. HbA1c is a less evenly reliable test across populations than its universal use implies. Different ancestry groups show slightly different HbA1c values at similar measured glucose, and researchers are still working out why.
When an HbA1c moves in a direction nothing else explains, ask what happened to the red cells before assuming something happened to the glucose. A recent bleed, six weeks of iron, a pregnancy, or a fall in kidney function can each shift the number without a single meal changing.
Two failure modes, then, both pointing the same way: the number needs company.
What is worth reading beside it?
Start with what HbA1c cannot see at all. A fasting insulin result reports the effort being spent to hold glucose down. That part of the story runs for years while glucose itself still reads normal. A glucose taken one to two hours after an ordinary meal catches the peaks the average smooths away. Where continuous monitoring is available, it shows what a single value never can: time spent in range, and how wide the daily swings actually are.
One more reading is already on most cholesterol panels: the ratio of triglycerides to HDL. Triglycerides are the blood’s transport fats, and HDL is often called the good cholesterol. Their ratio tracks how hard the body is working against insulin, closely enough to be worth a glance. Then comes the most underused reading of all, which costs nothing: the direction of travel. One HbA1c is a data point; three across four years, read beside a fasting insulin and a post-meal glucose, is information.
Proven: HbA1c reflects sugar attachment weighted towards recent weeks; that is straightforward biochemistry. Also proven: red cell lifespan shifts the value, documented in kidney disease, haemolysis, pregnancy and iron treatment. Promising: the idea that glucose swings add risk beyond the average, supported by mechanism and by observational data, with trials targeting the swings themselves still limited. The triglyceride to HDL ratio is a useful, inexpensive signal, though it performs unevenly across ancestry groups and should never stand alone. Being studied: why populations differ in HbA1c at similar glucose; researchers have not yet built a validated correction.
- HbA1c is a weighted figure, not a flat three-month average: about half comes from the last month.
- Identical averages can sit over very different days, and the swings carry risk the average hides.
- Shorter red cell survival lowers HbA1c independently of glucose; longer survival raises it.
- Haemoglobin variants interfere with some laboratory methods, so the test reads less evenly across populations.
- Fasting insulin, a post-meal glucose, time in range, the triglyceride to HDL ratio and the multi-year trend give the number its meaning.
Knowing what a test can and cannot see is the habit this whole field keeps returning to. The Metabolic Health knowledge check will show how firmly it has settled.
