On this page
- What the percentage is actually measuring
- 5.7 — the line itself
- 5.8 and 5.9
- 6.0 — the psychological line
- 6.1 and 6.2
- 6.3 and 6.4 — the edge
- 6.5 twice is diabetes — once is a flag
- Direction beats position
- When the number is not telling the truth
- What to do with your number this week
- Related
- Frequently asked questions
What the percentage is actually measuring
HbA1c is not this morning’s sugar. It is the fraction of your haemoglobin — the protein inside red blood cells — that has glucose stuck to it, and because a red cell lives about three months, the percentage reports an average of roughly ninety days of blood. That is why it cannot be improved by fasting the night before, and why one wedding lunch cannot ruin it. Between 5.7 and 6.4 percent the laboratory calls it prediabetes; every family in this district calls it borderline sugar, and there is no settled word for the stage in either language — diabetes itself is sugar noi in Tamil and prameham in Malayalam, but the stage before it goes unnamed, which is part of why it goes unheeded. The mechanics, including the conversion from percentage to average glucose, are set out on the HbA1c test explained page; what the whole band means as a diagnosis belongs to the borderline sugar guide. This page takes the values one at a time, because that is how people actually read their reports.
5.7 — the line itself
5.7 is not a rounding error, and it is not a diagnosis. It is the first percentage at which the laboratory stops calling you normal — the floor of the borderline band, an average sugar of roughly 117 mg/dL across three months. People who arrive at 5.7 usually feel slightly wronged by it, because 5.6 was fine and nothing about the body changed overnight. Nothing did. Thresholds are lines drawn across a slope so that doctors and patients have somewhere to stand and talk. What 5.7 buys is time and information: the pancreas is working harder than it should, and this is the earliest and cheapest point at which ordinary effort — a daily walk, a smaller mound of rice, four or five kilos — reliably moves the number back down. Nobody at 5.7 needs to panic. Plenty of people at 5.7 file the report away and meet it again as 6.8, four years later.
5.8 and 5.9
5.8 and 5.9 are the most anxiously searched numbers in the whole band, because they sit close enough to six to feel like a countdown. They are not a countdown. They are the lower half of a range whose message does not change from one end to the other. What is worth knowing is that risk rises steadily as you move through the band rather than jumping at any particular decimal — so 5.9 is somewhat further along than 5.7, somewhat less far along than 6.3, and none of those differences justify losing sleep over the third digit. The question that actually matters at 5.9 is what your last report said. A 5.9 that was 5.5 two years ago describes a body drifting in one direction with momentum behind it. A 5.9 that was 6.3 last year describes work that is succeeding. The same decimal, two entirely different conversations in the consulting room.
6.0 — the psychological line
6.0 is a psychological line, not a biological one. Nothing shifts inside the body between 5.9 and 6.0; what shifts is the reader, because a whole number feels like an event. Clinically, 6.0 sits in the middle of the borderline band — further from normal than anyone would like, still a comfortable distance below the diabetes threshold, and still entirely capable of coming back down. In practice 6.0 is a useful number precisely because it frightens people usefully: high enough to interrupt years of postponement, low enough that the interruption still works. If a routine test has just handed you 6.0, the correct response is not a kitchen full of imported flours and a subscription programme. It is a repeat test in three months, a measuring tape around the waist, and an honest look at what the last three months of walking and eating actually were.
6.1 and 6.2
6.1 and 6.2 belong to the upper half of the borderline band, and that is the only thing separating them from the values below — still prediabetes, still reversible, still short of a diagnosis. What changes in the upper half is the margin for delay. Someone at 6.2 has less room between where they stand and the threshold than someone at 5.8, so the same slow drift arrives sooner. This is also the part of the band where a good consultation stops looking only at the sugar number and starts looking at the blood pressure, the lipid report, the waist and the family history, because the borderline stage usually travels with company, and finding it early is only useful if the whole picture gets examined. If your report reads 6.1 or 6.2, ask for the rest of that picture rather than only a retest date.
6.3 and 6.4 — the edge
6.3 and 6.4 are still prediabetes — the last two values before the diabetes threshold, not the first two values of diabetes. That distinction is worth defending, because people at 6.4 frequently decide they have effectively been diagnosed already and respond by either giving up or treating themselves, and both are mistakes. At the same time, honesty requires saying that the room here has become small: without change, most reports sitting at 6.4 are travelling in a direction that crosses the line within a year or two. Two things belong in your hands at this end of the band. First, a repeat test done properly, because 6.4 is close enough to the threshold that laboratory-to-laboratory variation genuinely matters. Second, a consultation, because at 6.3 and 6.4 the assessment is worth doing with your fasting glucose, waist and history read together — not a percentage read alone at midnight on a phone.
6.5 twice is diabetes — once is a flag
6.5 percent is the diabetes threshold, and a diagnosis is made of two readings at or above it, or of one alongside a clearly abnormal glucose test or unmistakable symptoms. A single 6.5 on a routine screening, in someone who feels perfectly well, is a flag rather than a sentence. Repeat it. Laboratories differ, methods differ, and a recent illness, blood loss or a course of steroids can push a reading somewhere it does not belong. What a lone high number should never do is start, stop or adjust any medicine on its own authority — that decision belongs to a consultation with the full picture in front of it. What it should do is bring you in promptly, because if a second test confirms diabetes, the earliest years after diagnosis are the ones in which remission is most achievable, and they are also the easiest years to waste.
Direction beats position
Within this band, the direction of travel tells you more than the exact position in it. HbA1c averages three months, so two tests six months apart describe a trend that no single test can: 6.2 falling to 5.9 is a body responding, and 5.7 climbing to 6.0 is a body drifting, whatever the individual decimals look like on paper. The Indian evidence says the same thing at scale. ICMR-INDIAB, published in 2023, counted about 136 million Indians in the prediabetes zone alongside 101 million with diabetes. An Indian trial found that more than half of untreated borderline cases progressed within three years, while sustained lifestyle change cut new diabetes by nearly a third, and a Kerala cohort found roughly six in ten borderline cases crossing over within ten years. Those figures describe slopes, not decimals. If you have never been tested at all and landed here worrying, start with the 60-second risk score — age, waist, family history, activity, no needle involved.
When the number is not telling the truth
HbA1c rests on an assumption — red cells living a normal span of about 120 days — and wherever that assumption breaks, the percentage misleads. Anaemia and iron deficiency, both common in this region, can shift the value; recent blood loss, a transfusion or haemolysis can lower it falsely; advanced kidney disease distorts it in both directions; pregnancy changes red cell turnover enough that obstetric care uses direct glucose testing instead; and certain haemoglobin variants present in Indian populations interfere with some laboratory methods. So a borderline HbA1c is not the end of the enquiry. When the percentage and a fasting or post-meal reading disagree persistently, neither is discarded — the disagreement itself is the finding, and it is investigated. The blood sugar chart sets out the ranges those other tests are read against.
What to do with your number this week
Whatever your decimal, the instructions are nearly identical, which is the quiet point of this entire page. Repeat the test at three months — sooner mostly measures impatience, since the result cannot yet contain your effort. Measure the waist, because for South Asian bodies it says more than the weighing scale does: studies show we reach the diabetes risk of a heavyset European at a BMI of only 22, the fat sitting around the liver and abdomen rather than under the skin. Then start the two habits you can hold through this month — thirty minutes of walking most days, or ten minutes after each main meal, and a plate where the rice mound shrinks while vegetables, curd, fish or egg take back the space. The full method, and what the word reversal honestly means, is on how to reverse prediabetes. And if your number came from a relative’s glucometer or an online converter rather than a laboratory, one blood test settles it — book a test.
Related
Frequently asked questions
Not dangerous today, and not safe to ignore either. 5.9 sits in the lower half of the borderline band, where nothing has been damaged yet and normal regulation can still return. What decides its meaning is your previous report: 5.9 rising from 5.5 and 5.9 falling from 6.3 are opposite stories told by the same decimal.
No. Progression is common but not fixed. An Indian trial found that more than half of untreated borderline cases became diabetic within three years, and that sustained lifestyle change cut new diabetes by nearly a third. The percentage describes where you stand now; the next few years of weight, movement and testing decide the rest.
Three months, in most cases. HbA1c reflects roughly ninety days of blood, so a repeat at six weeks mostly measures impatience rather than progress. If your reading was 6.3 or 6.4, your doctor may want a fasting glucose sooner as well, because that end of the band sits close enough to the threshold for laboratory variation to matter.
Often both. Fasting glucose captures one morning; HbA1c captures three months, including every post-meal rise a fasting test sleeps through. A normal fasting reading with a borderline HbA1c commonly means sugar climbs after meals and settles by dawn. That is a reason to test after a meal too, not a reason to dismiss either result.
It can be misleading rather than wrong. Anaemia, iron deficiency, recent blood loss or transfusion, advanced kidney disease, pregnancy and certain haemoglobin variants all disturb the assumption the test rests on. If your percentage and your glucose readings keep disagreeing, say so at the consultation — that disagreement is a finding worth investigating, not an error to average away.