A routine blood test, and one line flagged. An HbA1c of 44 mmol/mol. Your GP mentions prediabetes, tells you to keep an eye on it and books the next test for a year's time. You leave not knowing whether it is a warning or the start of a disease.
Prediabetes usually causes no symptoms, and it is often picked up by chance. It can also go backwards, and that is what the biggest prevention trials showed.
What the numbers mean
In the UK, NICE talks about high risk, also called non-diabetic hyperglycaemia, with an HbA1c of 42 to 47 mmol/mol or a fasting blood sugar of 5.5 to 6.9 mmol/l (NICE). The American Diabetes Association sets the bar lower, from 39 mmol/mol and 5.6 mmol/l. From 48 mmol/mol or 7.0 mmol/l fasting, it is diabetes (ADA, 2025).
HbA1c reflects your average blood sugar over the last two to three months. A fasting test only tells you about that one morning. And risk does not flip at a threshold. It rises steadily, and faster at the top of the range.
What the big trials showed
In Finland, 522 overweight adults aged around 55, with high blood sugar after a glucose drink, were randomised. One group got one-to-one advice to lose weight, move more and change how they ate. They lost 4.2 kg on average in the first year. After four years, 11% of them had diabetes, against 23% in the control group, a 58% lower risk (Tuomilehto et al., 2001).
In the United States, the Diabetes Prevention Program followed 3,234 people for 2.8 years on average. The lifestyle group aimed to lose 7% of body weight and do at least 150 minutes of physical activity a week. It cut new cases of diabetes by 58% compared with placebo. Metformin, a diabetes medicine, did less well, at 31% (Knowler et al., 2002).
The effect lasted. Over seven years of follow-up in Finland, including three after the advice stopped, risk was still 43% lower, and 36% lower over the period with no support at all (Lindström et al., 2006).
Getting back to normal, even once
The American researchers then followed 1,990 participants for years. Those whose blood sugar had returned to normal at least once during the trial went on to have a 56% lower risk of diabetes than those who stayed prediabetic, whatever group they had been in (Perreault et al., 2012).
The reverse is also true. Participants who stayed prediabetic despite the intensive programme were the highest-risk group. What counts is actually bringing the number down, not just following a programme.
What mattered most
In the American trial, weight loss was the strongest factor. Each kilo lost cut the risk by 16%. Among participants who had not reached the weight goal after a year, those who were active enough still had 44% fewer cases of diabetes (Hamman et al., 2006).
The dietary advice of the time was mostly about less fat and more fibre. That is not what I recommend. Fibre mainly acts as a brake when the plate is full of processed food that sends blood sugar up, and the simpler move is to take that food away.
I rely on protein and fat at every meal, because they slow the flow of sugar into the blood. Half the plate vegetables, a quarter protein, a quarter carbs, with rice, potatoes, sweet potatoes or fruit rather than everyday bread, pasta and anything that comes out of a packet. Vegetables and protein first, carbs last. In fifteen people with prediabetes, that order cut blood sugar peaks by more than 40%, and I have gone through it in PMOS and diet.
Sleep, muscle and walking
In a large American study on sleep and heart health, 1,486 adults aged 53 to 93 reported how long they usually slept. Those sleeping 5 hours or less had 2.5 times the odds of having diabetes compared with those sleeping 7 to 8 hours, even without insomnia (Gottlieb et al., 2005). It is a snapshot, and on its own it does not prove that short sleep causes diabetes. With short nights night after night, though, it is very hard to shift your blood sugar. Daylight within an hour of waking and fewer bright screens in the evening help your body clock, which also sets how sensitive you are to insulin. Keeping your meals within a ten to twelve hour window and not eating late points the same way.
Muscle takes up most of the sugar from a meal. In a US national survey of 13,644 adults, every 10% rise in the share of muscle in body weight went with 11% less insulin resistance and 12% less prediabetes or diabetes (Srikanthan and Karlamangla, 2011). Two to three full-body strength sessions a week, with loads that go up bit by bit, are enough to build it.
Walking works even faster. A team at George Washington University followed ten people over 60, with fasting blood sugar between 5.8 and 6.9 mmol/l, who walked for fifteen minutes after each of their three meals, or for 45 minutes in one go in the morning or afternoon. The three short walks did at least as well on 24-hour blood sugar, and better than both others over the three hours after dinner (DiPietro et al., 2013). Your leg muscles pull sugar straight in, without waiting for insulin.
Medical follow-up
Prediabetes is monitored with a blood test. NICE recommends one at least once a year, ideally the same test each time so the results can be compared. Ask your GP for this, with an HbA1c. If treatment is offered, that is a decision you make with your doctor.
An HbA1c of 44 mmol/mol today can drop below 42 within a year. Start with what works from your very next meal, vegetables and protein before the carbs, and fifteen minutes of walking after dinner.
This article is for information only and does not replace medical advice.