metaboliq
HbA1c is the number people photograph and send to a family group. It is a genuinely good test and it is also the one most often read as a verdict, largely because a decimal point looks precise in a way the underlying biology is not.
HbA1c estimates your average blood glucose over roughly the previous two to three months. In the criteria most Indian laboratories print, below 5.7 per cent is normal. From 5.7 to 6.4 per cent is the band described as increased risk. And 6.5 per cent or above meets the diabetes threshold, once confirmed. So 5.9 sits in the middle band. That is a reason for a fuller look, not a diagnosis.
What does HbA1c actually measure?
Glucose in the blood attaches to haemoglobin inside red blood cells, and it stays attached for the life of the cell. Because red cells live around three months, the proportion of haemoglobin carrying glucose gives an average exposure over roughly the preceding two to three months, weighted towards the more recent weeks. That is the whole idea: one sample, no fasting, and a picture of a season rather than a morning.
It is also why it behaves differently from a fasting glucose. A single indulgent week barely moves it. A steady change over two months does. And anything that alters how long red cells live alters the result for reasons that have nothing to do with sugar at all.
Sources: [2]
What are the HbA1c bands?
In the criteria most Indian laboratories print, a value at or above 6.5 per cent meets the diagnostic threshold for diabetes once confirmed. Values in the 5.7 to 6.4 per cent band are described as indicating increased risk. Below that is reported as normal. Those are the thresholds set out in the standards of care used across most of clinical practice.
The World Health Organization's consultation on the test took a narrower position on the intermediate range. It recommended the diagnostic threshold, but declined to endorse a single lower cut-off for the at-risk band, on the grounds that the evidence for one was not strong enough. So a value of 5.9 sits inside a band that one major body defines and another explicitly declined to define. That is worth knowing before treating it as a line that has been crossed.
This is a real disagreement between guideline bodies, not a rounding difference. The bands are a useful convention laid over a continuous risk gradient, and the gradient has no step in it.
Is one HbA1c value a diagnosis?
Diagnostic criteria require confirmation. A single result in the diabetic range is repeated, or corroborated by a second different test, before the diagnosis is made. The exception is someone clearly unwell with symptoms, where the picture is not ambiguous. That requirement exists because assays vary, samples vary, and people vary week to week.
The same logic applies more strongly in the intermediate band, where a change of a tenth or two can move a report across a printed line without anything having changed in the person. Movement across two or three measurements in the same laboratory is informative. A single decimal place is not.
Sources: [1]
When is HbA1c the wrong test?
This is the part most often missed on a report, and it matters a lot in India. Anything that shortens red cell survival lowers the result, and anything that lengthens it raises the result — independently of glucose. Iron deficiency anaemia, recent blood loss, recent transfusion, pregnancy, chronic kidney or liver disease, and inherited haemoglobin variants all interfere.
Haemoglobin variants are not rare here. In populations where thalassaemia trait and sickle cell trait occur at appreciable frequency, an HbA1c can be misleading in either direction, depending on the variant and the assay. Guidance is explicit on this: the test should not be used diagnostically where such a condition is known or suspected. A doctor reading a report checks the haemoglobin and the red cell indices sitting next to it for exactly this reason.
Sources: [2]
The intermediate band is not a sentence
Reviews of the intermediate states are clear that movement occurs in both directions. A meaningful proportion of people in an at-risk band return to normal values on retesting, some remain stable for years, and some progress. Which of those happens is not readable from the value itself, and any page presenting the band as a countdown is describing a population trend as if it were an individual's future.
What is reasonably well established is that the intermediate state is a moment for attention rather than alarm. Long-term follow-up of trials of diet and activity changes in people with impaired glucose handling has reported delayed onset of diabetes over decades. That is an argument for taking it seriously and a poor argument for panic.
What to do with the number in your hand
Look at what surrounds it. A fasting glucose, a full blood count including haemoglobin, a lipid panel, liver markers and thyroid function are the context that turns one decimal into a description. In India's national survey, prediabetes, abdominal obesity and lipid abnormalities were found together often enough that reading any one of them alone is a decision to ignore the others.
Then get it read by somebody accountable. A number without a person attached to it produces one of two useless outcomes: dismissal, or a fortnight of searching that ends somewhere selling a supplement.
Sources: [4]