The report comes back with HbA1c 5.9, highlighted in yellow. The doctor says it is borderline, keep an eye on it, cut down on sweets. Nobody says what “keep an eye on it” means, how long you have, or what would actually move it. So you cut visible sugar for three weeks, feel virtuous, and retest a year later at 6.1.
Prediabetes is the most actionable finding in preventive medicine and the most weakly acted upon. The intervention that works is known in unusual detail — the size of the effect, the amount of weight, the minutes of exercise per week — because it has been tested in large randomised trials, including one done in India. Very little of that detail survives the thirty seconds it gets in a consultation.
The numbers worth carrying
First, the numbers — and why two labs will disagree
There is no single global definition, which is a real source of confusion when people compare reports.
The American Diabetes Association defines prediabetes as fasting glucose 100–125 mg/dL, or HbA1c 5.7–6.4 percent, or a 2-hour value of 140–199 mg/dL on an oral glucose tolerance test. The WHO sets its fasting threshold higher, at 110–125 mg/dL, and international expert guidance commonly puts the HbA1c band at 6.0–6.4 percent.
Why two labs disagree about the same number
| Test | ADA definition | WHO / international |
|---|---|---|
| Fasting glucose | 100–125 mg/dL | 110–125 mg/dL |
| HbA1c | 5.7–6.4% | 6.0–6.4% |
| 2-hour OGTT | 140–199 mg/dL | 140–199 mg/dL |
An HbA1c of 5.8 is prediabetes under one definition and normal under the other, and both reports are correct. Risk rises continuously rather than switching on at a threshold.
So an HbA1c of 5.8 is prediabetes under one definition and normal under another, and both reports are correct. That matters less than it seems. Risk rises continuously across these ranges rather than switching on at a threshold, and the practical reading of 5.8 is the same either way: you are on a trajectory, and the trajectory is modifiable.
One caveat on the OGTT — it is the least convenient of the three and the most informative. A meaningful number of people, South Asians particularly, have a normal fasting glucose and a clearly abnormal 2-hour value. If your fasting sugar keeps coming back fine but risk factors are stacking up, that is the test worth asking about.
The Indian version happens earlier, thinner and faster
National survey data puts roughly 15 percent of Indian adults in the prediabetes range, alongside about 11 percent with diabetes — and in several states the prediabetes number exceeds the diabetes number, which is a description of what is coming rather than what has happened.
The important part is that the risk arrives at a lower body weight. The South Asian phenotype tends toward less muscle and more visceral fat at any given BMI — often described as thin-fat — along with earlier decline in insulin secretion. Which is why the cutoffs are different and should be used as such: overweight starts at a BMI of 23 rather than 25, obesity at 25 rather than 30, and the waist thresholds are 90 cm for men and 80 cm for women.
The cutoffs that apply to you
| Measure | Standard cutoff | Asian Indian cutoff |
|---|---|---|
| Overweight | BMI 25 | BMI 23 |
| Obesity | BMI 30 | BMI 25 |
| Waist, men | — | 90 cm |
| Waist, women | — | 80 cm |
Less muscle and more visceral fat at any given BMI. The tape tells you more than the scale, because it proxies where the fat is.
The practical consequence: a 72 kg man with a BMI of 24 and a 94 cm waist who has been told his whole life that he is not overweight can be squarely in the risk group. Waist tells you more here than the scale does, because it is a proxy for where the fat is, and where the fat is — particularly in and around the liver — is what drives insulin resistance. This is the same mechanism that produces fatty liver, and the two findings usually travel together on the same report.
HbA1c can lie to you, in both directions
This gets almost no airtime and it changes how a report should be read. HbA1c measures glycated haemoglobin, and it assumes your red cells are behaving normally. When they are not, the number drifts — without your blood sugar having changed at all.
- Iron deficiency raises HbA1c. Long-lived, iron-starved red cells accumulate more glucose. A woman with heavy periods and low ferritin can be labelled prediabetic on a number her iron status inflated. It is worth reading alongside a ferritin result — and if you are already supplementing, there are several common reasons iron tablets do not work.
- Anything that shortens red cell lifespan lowers it — haemolysis, recent blood loss, a recent transfusion, some anaemias. Here the risk is the opposite one: falsely reassuring.
- Haemoglobin variants — and thalassaemia trait is not rare in parts of India — can interfere with certain assays.
- Chronic kidney disease and advanced liver disease both distort it.
When HbA1c misleads
Reads falsely high
- Iron deficiency. Long-lived, iron-starved red cells accumulate more glucose
- A woman with heavy periods and low ferritin can be labelled prediabetic on a number her iron status inflated
Reads falsely low
- Haemolysis
- Recent blood loss or a recent transfusion
- Some anaemias
- Here the risk is the opposite one — false reassurance
Also distorting: haemoglobin variants such as thalassaemia trait, which interfere with certain assays, plus chronic kidney disease and advanced liver disease. Read a borderline HbA1c next to a haemogram and a ferritin.
None of this makes HbA1c a bad test. It makes it a test that needs a haemogram and a ferritin next to it before anyone acts on a borderline value.
What actually reverses it, with the numbers attached
The Diabetes Prevention Program randomised over three thousand adults with impaired glucose tolerance to intensive lifestyle change, metformin, or placebo. The lifestyle arm had two targets: 7 percent body weight loss and 150 minutes a week of moderate activity.
Over about three years, that arm reduced progression to type 2 diabetes by 58 percent. Metformin reduced it by 31 percent. In participants over 60, lifestyle change reduced it by 71 percent. Lifestyle beat the drug, decisively, and it beat it hardest in the oldest group — which is the reverse of what most people assume.
The Indian Diabetes Prevention Programme ran the same question in an Indian population with impaired glucose tolerance, and the pattern held with a local twist: lifestyle modification reduced progression by around 28 percent, metformin by a similar margin, and — notably — combining them added nothing over either alone. The absolute risk being prevented was larger, because Indian participants progressed faster to begin with.
Reduction in progression to type 2 diabetes
Bars scaled to the largest effect. In the Indian DPP, metformin achieved a similar margin to lifestyle change — and combining the two added nothing over either alone. The absolute risk prevented was larger, because Indian participants progressed faster to begin with.
Two things to take from this. The intervention is specific: 7 percent and 150 minutes, not “eat healthy and stay active”. And these are the effects of change that was sustained — the trials measured years, not a January.
Aim for normal, not for “not worse”
The default goal people are given is to avoid progressing. The better goal, and a genuinely achievable one, is to get back to normal glucose regulation.
The distinction is not academic. Follow-up of the DPP cohort found that people who regressed to normal glucose — even briefly at some point during the study — had substantially lower long-term risk of developing diabetes than those who merely stayed prediabetic without progressing. Getting the number back into the normal range appears to reset something meaningful, and it is worth aiming at explicitly rather than settling for stability.
The levers, in the order I would pull them
1. Weight, but specifically waist. Seven percent is the trial target, and it is a smaller number than it sounds — 5.6 kg for an 80 kg person. Track waist alongside weight, because visceral fat responds early and the tape often moves before the scale does.
2. Muscle, because it is where glucose goes. Skeletal muscle is the largest site of glucose disposal in the body, and a single resistance session improves insulin sensitivity for a day or more afterwards. Building muscle enlarges the tank permanently. For the thin-fat phenotype this is the highest-leverage change available, and it is the one most often skipped in favour of walking. Walk as well — but start with the weights.
3. What surrounds the carbohydrate, more than the carbohydrate itself. Nobody in India is giving up rice or roti, and the plan that demands it fails in week three. What works is changing the rest of the plate: protein and vegetables in meaningful quantity at the same meal, whole pulses left intact, and the largest carbohydrate portion moved to the meal followed by activity rather than by sleep. Same staple, materially different glucose response.
4. Liquid sugar and refined snacks, which are the easy win. Sweetened drinks, packaged juice, biscuits with tea twice a day. This is usually the first thing people cut and, on its own, rarely enough — but it costs nothing to do.
5. Sleep and daily steps. A week of short sleep measurably worsens insulin sensitivity in healthy adults. Steps outside your workout do more for glucose control across a day than the workout does. Both are unglamorous and both are usually where the plan is actually failing.
On fasting protocols: intermittent fasting can work well for glucose control in the right person, and poorly in someone who compresses the same intake into a smaller window and sleeps badly for it. If you are considering it, the practical version is here. On metformin: it is a reasonable option in specific higher-risk situations and that is a decision for your doctor — if you do end up on it long term, be aware it depletes vitamin B12, which almost nobody checks.
What to track, and how often
HbA1c every six months while you are actively changing things, not every six weeks — it reflects roughly three months of glucose and testing it early tells you nothing. Alongside it: fasting glucose, a fasting lipid profile, ALT for the liver, blood pressure, ferritin and a haemogram so the HbA1c can be interpreted properly, and a waist measurement taken the same way each time. If you have PCOS, insulin resistance is likely the driver of both problems and is worth treating as the primary target.
What to track, and how often
| What | How often | Why |
|---|---|---|
| HbA1c | Every 6 months | It reflects roughly three months of glucose. Testing at six weeks tells you nothing useful |
| Fasting glucose | With each HbA1c | Catches the fasting pattern the HbA1c averages away |
| Ferritin + haemogram | At least once | So the HbA1c can be interpreted properly |
| Fasting lipids | 6–12 months | Triglycerides and HDL move with insulin resistance |
| ALT | 6–12 months | Fatty liver travels with this on the same report |
| Waist | Monthly | Same method each time. Visceral fat responds early, so the tape often moves before the scale |
What to take away
- Definitions differ between the ADA and the WHO, so two labs can disagree about the same number. Risk is a gradient, not a switch
- Indians develop this earlier and at lower body weight. Use BMI 23 and waist 90 cm / 80 cm, and trust the tape over the scale
- Iron deficiency inflates HbA1c. Read a borderline result alongside ferritin before accepting the label
- The trial-proven target is 7 percent body weight and 150 minutes a week — a 58 percent reduction in progression, more than metformin achieved
- Aim to get back to normal glucose, not merely to avoid getting worse. Regression carries lasting benefit
- Resistance training is the highest-return change for the South Asian phenotype, because muscle is where glucose is disposed of
- Recheck HbA1c at six months, not six weeks
Before you act on any of this
I am a pharmacist and a nutritionist. I am not your doctor, and this is general information rather than advice about your results. If you already have type 2 diabetes, are on glucose-lowering medication, or are pregnant — where glucose targets and screening are different and stricter — your management belongs with your physician, and changing your eating or activity substantially while on insulin or a sulfonylurea needs your doses reviewed first, because the risk is hypoglycaemia. Nothing here is a reason to start, stop or change a prescribed medicine, including metformin. What this article can do is tell you which numbers are worth asking about and what size of change the evidence actually supports. If you want that built into a plan you will still be following in month six, start with a one-time assessment.
