Prediabetes: The Ten-Year Warning Most People Miss

Most people meet the word "prediabetes" on a printed result, in a sentence that ends with "come back in a year". It sounds like nothing much. In fact it is the most informative thing a routine blood test will ever tell you about the next twenty years of your health, and the biology behind it started long before the number moved.

What insulin resistance actually is#

Insulin has one main job in this context: it tells muscle and fat cells to take glucose out of the blood, and it tells the liver to stop releasing its own stored glucose.

Insulin resistance means those tissues have become less responsive to that signal. Not deaf: just harder of hearing. The pancreas responds the way anyone would when they are not being heard: it raises its voice. Insulin output goes up, sometimes to three or four times normal, and blood glucose stays perfectly normal.

This is why you can be insulin resistant for a decade or more with flawless blood sugar results. The compensation is doing its job. Nothing looks wrong on a standard test.

The compensation eventually falters. Beta cells, the insulin-producing cells in the pancreas, are working under sustained strain, and in people genetically susceptible they begin to fail. Glucose starts to drift upward. First after meals, then in the fasting state. That drift is what we label prediabetes. By the time it appears, studies of people followed for years suggest beta-cell function is often already reduced by roughly half.

Why the fat matters more than the weight#

The strongest driver is not fat under the skin. It is fat in the wrong places: inside liver cells, inside muscle, and around the abdominal organs. Fat inside the liver makes the liver ignore insulin's instruction to stop producing glucose, which is why fasting glucose rises overnight. Fat inside the pancreas appears to impair insulin secretion directly.

Each person seems to have a personal fat threshold. An amount of fat their subcutaneous tissue can comfortably store. Cross it, and the surplus spills into liver and muscle. Some people cross that threshold at a BMI of 35. Others cross it at 22.

The numbers, in both units#

Three different tests, three different sets of thresholds, and two organizations that do not fully agree. Here is the whole picture.

A diagnosis of diabetes normally needs either two abnormal results on separate days, or one abnormal result plus clear symptoms.

When HbA1c lies#

HbA1c measures how much glucose has stuck to hemoglobin over the roughly three-month lifespan of a red blood cell. Anything that changes that lifespan changes the number without changing the blood sugar.

• Falsely low: iron deficiency treatment in progress, recent blood loss or transfusion, haemolysis, pregnancy, advanced liver disease, and some hemoglobin variants.
• Falsely high: untreated iron deficiency anemia, vitamin B12 deficiency, splenectomy, and chronic kidney disease.
• Unreliable altogether: sickle cell disease, beta thalassaemia and other hemoglobinopathies, which matter enormously in South Asia, the Mediterranean, the Gulf and sub-Saharan Africa, and which are precisely the regions where diabetes is rising fastest.

If you have any of these, fasting glucose or an oral glucose tolerance test is the more honest measurement, and it is reasonable to ask for it.

Why South Asians develop this at a lower BMI#

This is not a minor adjustment. South Asians develop type 2 diabetes roughly five to ten years earlier and at a body mass index about 3 to 5 units lower than white European populations, and the risk is two to four times higher at any given BMI.

Several things stack:

• Body composition. At the same BMI, South Asians carry more total body fat, more visceral fat around the organs, more liver fat, and less skeletal muscle. Muscle is where most glucose is disposed of after a meal, so less of it means less capacity.
• Where fat is stored. Subcutaneous fat in the lower body appears less expandable, so surplus energy reaches the liver and viscera earlier. The so-called thin-fat phenotype, visible even in newborns.
• Beta-cell reserve. Insulin secretory capacity appears lower on average, so the compensation phase runs out sooner.
• Early-life programming. Low birth weight followed by rapid childhood weight gain is common across the region and is associated with adult insulin resistance.
• Rapid environmental change. A generation-scale shift towards refined carbohydrate, sedentary work and urban living, layered on a genome shaped by very different conditions.

The practical consequence is in the cut-offs. For people of South Asian, Chinese, other Asian, Middle Eastern, Black African and African-Caribbean ancestry, the overweight threshold is a BMI of 23 and the obesity threshold is 27.5, not 25 and 30. Waist circumference thresholds are lower too: 90 cm for South Asian men and 80 cm for South Asian women, compared with 94 cm and 80 cm in European reference data.

If a clinician has told you your BMI of 24 is fine, and you are South Asian with a family history of diabetes, that is a reasonable thing to question politely.

What the prevention trials actually showed#

This is the part that deserves more attention than it gets, because the results are unusually strong for lifestyle research.

The Diabetes Prevention Program (DPP), United States, 3,234 adults with impaired glucose tolerance, average follow-up 2.8 years. Three arms: intensive lifestyle program, metformin, or placebo. The lifestyle arm aimed for a 7 per cent reduction in body weight and 150 minutes of activity a week, with sixteen structured sessions and ongoing coaching.

• Lifestyle program: 58 per cent lower incidence of diabetes than placebo.
• Metformin: 31 per cent lower.
• In participants aged 60 and over, the lifestyle arm reached 71 per cent, while metformin was barely better than placebo in that group.
• Within the lifestyle arm, each kilogram of weight lost was associated with about a 16 per cent lower risk.

The Finnish Diabetes Prevention Study, 522 participants, found the same 58 per cent reduction using individualised dietary and exercise counseling, and the benefit persisted for years after the active intervention stopped.

The Da Qing study, China, is the long one. Six years of intervention, then thirty years of follow-up. At the thirty-year mark the intervention group had a delay in diabetes onset of about four years, fewer cardiovascular events, fewer microvascular complications, lower cardiovascular death, and roughly 1.4 additional years of life expectancy. Six years of effort, three decades of return.

The Indian Diabetes Prevention Program (IDPP-1) matters for the global picture. In a South Asian population that was younger and leaner than the American cohort, lifestyle modification reduced progression by about 29 per cent. A real effect, but a smaller one, and a reminder that the same intervention does not transfer identically across populations.

The DPP Outcomes Study, following the original participants for 15 years, found the lifestyle benefit had narrowed to 27 per cent and metformin to 18 per cent, largely because the control group eventually received advice too. The gap narrowed; it did not close.

What genuinely helps, and what is oversold#

Well supported. Regular aerobic activity, including brisk walking. Resistance training, which builds the muscle that disposes of glucose. Walking for ten minutes after meals, which blunts the post-meal glucose rise. Improving sleep duration and regularity, even a few nights of short sleep measurably reduces insulin sensitivity in healthy volunteers. Reducing sugar-sweetened drinks. Fiber from pulses, vegetables and whole grains. Treating obstructive sleep apnoea, which is both common and strongly linked to insulin resistance.

Genuinely helpful for some, worth discussing rather than assuming. Structured weight management, where it is wanted and sustainable. Metformin for specific higher-risk groups. Treating PCOS-related insulin resistance.

Oversold. Cinnamon, chromium, bitter melon, apple cider vinegar and berberine all have small studies with modest or inconsistent effects, and berberine in particular is sold with claims far beyond its evidence and has real drug interactions. Detox regimens do nothing. Continuous glucose monitors in people without diabetes generate data, not outcomes. And any product promising to "reverse" diabetes in a fixed number of days is making a claim no trial supports.

What I get asked most#

The question is almost always some version of: "So do I have diabetes or not?"

The honest answer is that the body does not work in categories. Glucose regulation is a slope, and the thresholds are lines we have drawn across it for practical reasons. Someone at 6.3 per cent and someone at 6.6 per cent are in different diagnostic boxes and have almost identical biology.

The second thing I find myself repeating: people fixate on the number and ignore the trajectory. An HbA1c that has gone 5.4, 5.7, 6.0 over three years tells me far more than a single 6.1. Ask for your previous results. The direction is the diagnosis.

And I have lost count of the number of people, mostly men in their thirties and forties, who were told years ago that they were "borderline", understood that to mean "fine", and returned with a reading well past the threshold. The word borderline may be the most expensive word in clinical medicine.

At work#

Shift work is an independent risk factor for type 2 diabetes, with the risk rising as the number of night shifts accumulates. The mechanism is circadian misalignment: eating at a time when the body's insulin sensitivity is at its lowest, combined with disrupted sleep, which itself reduces insulin sensitivity within days.

In an industrial setting the practical levers are usually not medical. They are the vending machine and the night canteen. Where the only food available at 3 a.m. is fried and sweet, glucose control is a facilities decision as much as an individual one. Rotating shifts forward, mornings to afternoons to nights, is tolerated better than rotating backwards.

Workplace health screening is also where a great deal of prediabetes is found, because it tests people who feel entirely well. If your employer offers an annual HbA1c, take it, and keep your own copy of the result so you can see the trend.

When to get help now#

Go to an emergency department for vomiting with abdominal pain, deep rapid breathing, drowsiness or a sweet, fruity smell on the breath. These suggest a metabolic emergency rather than gradual drift.

Seek prompt medical review for marked thirst, passing large volumes of urine, blurred vision, unexplained weight loss, recurrent thrush or skin infections, or new numbness and burning in both feet. Those are symptoms of established high blood sugar, not of prediabetes.

Book a routine test if you are over 40, or over 25 with South Asian, Chinese, Black African or African-Caribbean ancestry, if a parent or sibling has type 2 diabetes, if you have had gestational diabetes or PCOS, or if you have never been tested and cannot remember the last time anyone checked.

The bottom line#

Prediabetes is not a mild disease; it is the visible end of a process that has been running quietly for a decade. The thresholds differ between the US and the rest of the world, so read your result alongside the reference range on the report and, more importantly, alongside your previous results. If you are of South Asian or other Asian, Middle Eastern or African ancestry, the risk arrives at a lower body weight than the standard charts suggest. And the prevention trials are genuinely encouraging: structured, supported lifestyle change cut progression by more than half, worked best in the people most often written off as too old to bother, and was still paying out thirty years later.