The plan

How do I reverse prediabetes?

By GyroscopeUpdated November 14, 202613 sources

Lose 5 to 7% of your weight if you carry extra, walk briskly 150 minutes a week with 10 minutes after each meal, and retest your A1c at the same lab 12 weeks later. In the biggest trial, that plan cut new cases of diabetes by 58%, almost twice as much as the drug metformin. Work out your own numbers and dates below.

Why does prediabetes advice feel like a mess?

You get a lab result with an A1c of 5.9% and a note that says “prediabetes, watch your diet”. Then the internet takes over. Cut all carbs. Drink apple cider vinegar. Try a supplement. Buy a sensor. Fast until noon. Everyone has a trick, and almost none of them come with a number.

The plan with the strongest evidence is older and more boring. It was tested on more than 3,000 people for almost three years, it has a clear goal and a clear finish line, and it doesn’t care which diet you pick.

What’s the plan, week by week?

1. Week 0: get your numbers. A lab A1c and a fasting glucose, from the same draw if you can. Weigh yourself in the morning and write it down. Work out your weight target and retest date with the calculator below.
2. Week 1: start walking after dinner. Ten minutes, as soon as you finish eating. Dinner is the meal most people sit after, and in the walking trial it’s where a walk made the biggest difference, 22%.
3. Weeks 1 to 2: add the other meals. Ten minutes after each of three meals is 30 a day, so you reach 150 minutes a week by day 5. Walk briskly enough that you can talk but not sing.
4. Weeks 1 to 12: lose 5 to 7% if you carry extra. At 200 pounds that’s 10 to 14 pounds, about 1 pound a week. Eat a little less than you burn, with protein at every meal so you keep your muscle. tdee.guide works out your calories to lose, and bodyfat.guide your protein.
5. Every meal: carbs last. Protein and vegetables first, the bread, rice or pasta about 10 minutes later. In 15 people with prediabetes, that order cut the peak by about 46%.
6. Every night: 7 hours. A week of 5-hour nights made healthy men’s insulin work 11% worse.
7. Week 12: retest at the same lab. A1c and fasting glucose again. Then test once a year, as the ADA asks for anyone who’s been in the prediabetes range.

What did the trial actually do?

The Diabetes Prevention Program enrolled 3,234 American adults between 1996 and 1999. All had a 2-hour number of 140 to 199 mg/dL after the lab’s glucose drink, which is the prediabetes range, and a fasting glucose of 95 to 125. Their average age was about 51 and their average BMI 34. They were split three ways.

The lifestyle group had a 16-lesson course with a coach over the first 24 weeks, then monthly sessions. The goals were to lose at least 7% of their weight on a healthy low-calorie, low-fat diet and to do at least 150 minutes a week of activity such as brisk walking. A second group took metformin, and a third took a placebo pill.

The trial was meant to run longer. It was stopped a year early, in May 2001, on the advice of its data monitoring board, because the result was already clear. Over an average of 2.8 years, the placebo group developed diabetes at 11.0 cases per 100 people per year. Metformin cut that to 7.8, and the lifestyle program to 4.8. That’s 58% fewer new cases with lifestyle and 31% fewer with metformin. About 7 people had to do the program for 3 years to prevent one case of diabetes.

Nobody in the lifestyle group was perfect. Half reached the 7% goal by 24 weeks, and 38% were still there at their last visit. On average the group lost about 5.6 kg (12 pounds). 74% met the activity goal at 24 weeks, and 58% at the last visit.

Which part of the plan did the work?

Weight did most of it. When researchers went back through 1,079 people in the lifestyle group over 3.2 years of follow-up, weight loss was the main thing that predicted who stayed free of diabetes. Each kilogram lost (2.2 pounds) went with 16% lower risk, even after adjusting for changes in diet and activity.

Activity still counted on its own. Among 495 people who missed the weight goal at the end of the first year, those who met the activity goal had 44% lower diabetes risk. If the scale stalls, keep walking.

The 16% comes from that trial’s people, and it doesn’t stack. Losing 10 kilograms doesn’t mean 160% lower risk.

Why does losing weight work so fast?

Fat stored in and around the liver seems to be a big part of it. In a 2011 study, 11 people with type 2 diabetes went on 600 calories a day under close supervision. Within one week their fasting blood sugar fell from 9.2 to 5.9 mmol/L (166 to 106 mg/dL). By week 8 their liver fat had dropped from 12.8% to 2.9%, and their pancreas was making insulin normally again. That diet is extreme and the study was tiny, so nobody should copy it without a doctor. It shows how quickly the liver answers once weight starts coming down. The 12-week plan runs at a normal pace.

Do I have to cut carbs?

No. The trial used a low-fat diet. The ADA’s 2026 standards say “there is no ideal percentage of calories from carbohydrate, protein, and fat” for preventing diabetes, and name Mediterranean and low-carb eating among the options. In the trial, weight lost predicted success even after adjusting for diet. Cut refined carbs and sugary drinks first, since they’re the easiest calories to drop, and keep the rest of your eating in a shape you can live with. More on whether you have to cut carbs.

Does it last?

Both are true. A 12-week push won’t protect you for life, and getting your numbers back to normal even once still cut later risk by more than half. Plan the second year the way you planned the first. The walks after meals and the yearly A1c are the parts that keep going. Each habit is sized on its own in the habits that cut your type 2 risk, and if you already have type 2, can type 2 diabetes be reversed? covers remission.

What happens at the 12-week retest?

A1c moves slowly. After blood sugar came down in people newly diagnosed with type 2, A1c fell with a half-time of about 35 days, so by 12 weeks you see about four fifths of a lasting change. A retest at 4 weeks mostly shows your old habits.

• Back under 5.7% and under 100 fasting. You’re where people in the trial had lower risk for years afterward. Keep the walks and test every year.
• Lower, but still in the prediabetes range. It’s working. Keep going for another 12 weeks, and look at whether weight or walking slipped.
• The same or higher. Check the plan against what you actually did, and talk to your doctor about what else could help, including metformin.
• In the diabetes range. Book the second test this week and see your doctor.

Who should do this with a doctor?

Plan it with your doctor or midwife if you’re pregnant, if you have a history of an eating disorder, or if you take insulin or a sulfonylurea, which can make blood sugar fall too low as you change how you eat and move.

Metformin is a prescription decision. The ADA suggests doctors consider it for some people with prediabetes, such as those aged 25 to 59 with a BMI of 35 or more, a fasting glucose of 110 or more, an A1c of 6.0% or more, or a history of diabetes in pregnancy. If you’re on a GLP-1 drug such as semaglutide or tirzepatide, your prescriber sets your targets, and bodyfat.guide explains how GLP-1 drugs work.

If your BMI is under 25 (23 with Asian descent), weight isn’t your main way in. Start with walking after meals and a full night’s sleep. Could I have prediabetes? builds a plan from your answers.