GLP-1 Drugs for Prediabetes: Trial Results on Preventing Type 2 Diabetes, and the Rebound After Stopping

In clinical trials, GLP-1 drugs sharply cut the number of people with prediabetes who went on to develop type 2 diabetes. The strongest result comes from tirzepatide, sold as Zepbound and Mounjaro. Over about three and a half years, 1.3% of people with obesity and prediabetes who took it were diagnosed with type 2 diabetes, compared with 13.3% of those on placebo.

Two limits matter as much as that number. First, no drug is approved by the FDA specifically to prevent type 2 diabetes, including GLP-1 drugs and metformin. Second, the benefit depends on staying on the drug. When people in the semaglutide and tirzepatide trials stopped, blood sugar and weight started moving back toward where they began.

This guide covers what the trials found, how GLP-1 drugs compare with lifestyle change and metformin, what a new Mayo Clinic study adds, and what it costs to use these drugs for this purpose in 2026.

Prediabetes means blood sugar is higher than normal but not high enough to count as diabetes. The CDC uses three tests to define it:

• A1C, which reflects average blood sugar over the previous two to three months: 5.7% to 6.4%, with diabetes starting at 6.5%.
• Fasting blood sugar, measured after an overnight fast: 100 to 125 mg/dL, with diabetes starting at 126 mg/dL.
• Two-hour glucose tolerance test, taken after drinking a sugary liquid: 140 to 199 mg/dL, with diabetes starting at 200 mg/dL.

The CDC estimates that 115.2 million U.S. adults have prediabetes, including 31.3 million people aged 65 or older, which is just over half of that age group. Most of them do not know it: the CDC says 8 in 10 people with prediabetes are unaware of their condition because it usually causes no symptoms.

The U.S. Preventive Services Task Force recommends screening adults aged 35 to 70 who have overweight or obesity, and referring people with prediabetes to prevention programs. That screening is how most people find out, and it is the point at which the question of medication first comes up.

Any new prevention drug gets compared with the Diabetes Prevention Program (DPP), a U.S. trial published in the New England Journal of Medicine in 2002. It enrolled 3,234 adults with high blood sugar who did not yet have diabetes. Their average BMI was 34.0. They were assigned to placebo, the diabetes drug metformin, or a lifestyle program aimed at losing at least 7% of body weight and exercising 150 minutes a week.

Over an average of 2.8 years, the lifestyle program cut new cases of diabetes by 58% compared with placebo, and metformin cut them by 31%. Put another way, about 7 people had to go through the lifestyle program, and about 14 had to take metformin, to prevent one case of diabetes over three years.

The effect lasted, though it got smaller. In a 21-year follow-up published in 2025, people first assigned to the lifestyle program still had a 24% lower rate of diabetes than the placebo group, and those first given metformin had a 17% lower rate. Most of the benefit came in the first three years.

The American Diabetes Association's 2026 Standards of Care point to weight loss as the main reason the program worked: in the DPP, each kilogram lost (about 2.2 pounds) lowered the risk of diabetes by 16%. That link between weight and diabetes risk is also why GLP-1 drugs, which cause more weight loss than diet alone for most people, were expected to work for prevention.

The longest GLP-1 prevention data come from SURMOUNT-1, an Eli Lilly trial of tirzepatide. Of the 2,539 adults with obesity in the trial, 1,032 also had prediabetes. That group stayed on weekly tirzepatide (5, 10, or 15 mg) or placebo for 176 weeks, then stopped for 17 weeks. The results were published in the New England Journal of Medicine in 2025.

At 176 weeks, people on tirzepatide had lost 12.3% to 19.7% of their body weight, depending on the dose, compared with 1.3% on placebo. Type 2 diabetes was diagnosed in 1.3% of the tirzepatide group and 13.3% of the placebo group, a hazard ratio of 0.07, or about 93% lower risk. In absolute terms the gap was 12 percentage points, which works out to roughly one diagnosis avoided for every eight people treated for about three and a half years.

The most common side effects, apart from COVID-19, were stomach-related. Most were mild to moderate and happened in the first 20 weeks, while the dose was being raised. The authors reported no new safety signals.

Novo Nordisk ran a trial built for this question, STEP 10, which enrolled 207 adults in Canada, Denmark, Finland, Spain, and the U.K. who had a BMI of 30 or higher and prediabetes. They took semaglutide 2.4 mg (the Wegovy dose) or placebo for 52 weeks. It used the U.K. definition of prediabetes, which starts at an A1C of 6.0% instead of the U.S. cutoff of 5.7%. Normal blood sugar meant an A1C under 6.0% and a fasting glucose under 5.5 mmol/L, about 99 mg/dL.

At one year, 81% of the semaglutide group had normal blood sugar by that definition, compared with 14% on placebo, and weight fell 13.9% versus 2.7%. Serious side effects happened in 9% of each group, but 6% of semaglutide users stopped the drug because of side effects, compared with 1% on placebo. The trial reported normal blood sugar, not new diabetes diagnoses, and one year is too short to measure prevention directly.

An earlier analysis pooled 1,536 people with prediabetes from the STEP 1, 3, and 4 obesity trials. After 68 weeks, more people on semaglutide had normal blood sugar than people on placebo; in STEP 1, for example, the figures were 84.1% and 47.8%. The placebo numbers are worth noticing, since many people in those groups also improved. Every arm of these trials included diet and exercise counseling, which may explain part of it.

Before semaglutide and tirzepatide, Novo Nordisk tested liraglutide 3.0 mg (sold as Saxenda) in 2,254 adults with prediabetes and either obesity or overweight plus another health problem. The results, published in The Lancet in 2017, covered 160 weeks.

During treatment, 2% of the liraglutide group and 6% of the placebo group were diagnosed with diabetes. The hazard ratio was 0.21, or about 79% lower risk. Weight fell 6.1% versus 1.9%. The trial had a large weakness: about half the participants dropped out (47% on liraglutide and 55% on placebo), and the researchers did not follow people after they left. Liraglutide is a daily shot, and it produced less weight loss in this trial than semaglutide and tirzepatide did in theirs.

Randomized trials compare a drug with placebo. Doctors also want to know how GLP-1 drugs compare with metformin, the usual first drug for prediabetes. On November 2, 2026, the American Heart Association released a study from Mayo Clinic in Arizona that looked at this question using health records.

The researchers used the TriNetX network of health records from 104 institutions between 2010 and 2023. They studied 256,492 adults with prediabetes and obesity. The average age was 52, and 76% were women. About half took a GLP-1 drug (semaglutide, tirzepatide, liraglutide, or dulaglutide) and half took metformin. The two groups were matched on age, other illnesses, and other medicines. Over about three years, people on GLP-1 drugs had a 31% lower risk of type 2 diabetes and a 41% lower risk of death from any cause. They were also 1.5 times more likely to see blood sugar return to normal and had fewer heart attacks, strokes, and heart failure cases.

This is a meeting abstract for the AHA's Scientific Sessions, held November 6 to 9 in Chicago. It has not been peer-reviewed, and because it is a records study, it can show a link but cannot prove the drugs caused the difference. People who get GLP-1 prescriptions may differ from those given metformin in ways the records do not capture, and the study could not see how consistently people took their medicine or at what dose. The lead author, research fellow Vinicius de Sousa Barzon Serra, said metformin "is still a first-line treatment choice. It is safe, inexpensive and has a long track record." Robert Eckel of the University of Colorado, a past AHA president who was not involved, said the benefits likely come from weight loss and that "costs and access to these medications are important factors."

It is tempting to line up 93% for tirzepatide, 79% for liraglutide, 58% for the lifestyle program, and 31% for metformin and call it a ranking. The studies do not allow that. They enrolled different people, used different definitions of prediabetes, lasted different lengths of time, and ran decades apart.

The placebo groups show the problem. In the DPP, which required both high fasting glucose and a high glucose tolerance result, the placebo group developed diabetes at 11 cases per 100 people each year. In SURMOUNT-1, 13.3% of the placebo group developed diabetes over more than three years. Those are very different starting risks. A percentage cut in risk means more for a group that starts at higher risk.

What the studies do agree on is direction. Losing weight lowers the chance that prediabetes becomes diabetes, and GLP-1 drugs cause more weight loss on average than lifestyle programs or metformin. The ADA notes that it remains unknown whether lowering blood sugar early, before it reaches the diabetes range, improves long-term health. No trial has yet followed GLP-1 users with prediabetes long enough to show fewer heart attacks, kidney failures, or deaths from the prevention itself.

Both large GLP-1 trials checked what happened after people stopped the drug, and in both, much of the benefit faded.

In SURMOUNT-1, after 17 weeks off tirzepatide, the share of people with a diabetes diagnosis rose from 1.3% to 2.4%, while on placebo it went from 13.3% to 13.7%. The tirzepatide group still came out well ahead, with a hazard ratio of 0.12, but 17 weeks is a short window.

The STEP 1 extension followed 327 people for a year after they stopped semaglutide. They regained about two-thirds of the weight they had lost. Among those who started with prediabetes, 93.6% on semaglutide had normal blood sugar at the end of treatment, compared with 41.5% on placebo. One year after stopping, it was 43.3% versus 34.0%. That subgroup was small, with 142 people on semaglutide and 53 on placebo.

The ADA's 2026 Standards say any drug used for prevention will likely need to be taken long-term "because of waning effects after stopping the medication." For someone with prediabetes, starting a GLP-1 drug may mean taking it for years. Our maintenance playbook covers what trials show about lower maintenance doses and switching to pills.

The ADA's 2026 Standards of Care state that "there are no pharmacologic agents approved by the U.S. Food and Drug Administration specifically for the prevention of type 2 diabetes." That includes metformin, which doctors prescribe off-label for this use.

The ADA recommends that doctors consider metformin for people at the highest risk, especially adults aged 25 to 59 with a BMI of 35 or more, higher fasting glucose (such as 110 mg/dL or above), and higher A1C (such as 6.0% or above), as well as women who had diabetes during pregnancy. On GLP-1 drugs, the ADA draws a line. It says there are no long-term data to support any drug other than metformin "for the sole purpose of preventing type 2 diabetes." But it also says GLP-1 drugs used for weight management in people with overweight or obesity are "highly beneficial and should be considered," since weight loss is central to prevention.

In practice, that means the prescription is for weight. Wegovy and Zepbound are FDA-approved for adults with obesity, or with overweight plus at least one weight-related health problem. Neither label lists prediabetes as its own indication. Ozempic and Mounjaro are approved for type 2 diabetes, so prescribing them for prediabetes alone is off-label. Our guide to what each GLP-1 drug is approved for lists every current indication.

Coverage depends on the plan and on whether you meet the weight-loss criteria.

Medicare. Medicare drug plans have long been barred from covering drugs for weight loss. The Medicare GLP-1 Bridge, a CMS program running from July 1, 2026, through December 31, 2027, makes an exception for some people with Part D coverage. Prediabetes is one of the conditions it names: you can qualify with a BMI of 27 or more plus prediabetes, a past heart attack, a past stroke, or symptomatic peripheral artery disease. People with a BMI of 35 or more qualify without another condition. The copay is $50 a month for the Wegovy pen or pill, the Zepbound KwikPen, or Foundayo. People who already have type 2 diabetes use their regular Part D coverage instead. Our Medicare Bridge guide explains how to apply. Medicare also pays for the Medicare Diabetes Prevention Program, a lifestyle program run by CDC-recognized groups that have enrolled as Medicare suppliers.

Private insurance. Many employer plans do not cover weight-loss drugs, and plans that do often require prior authorization. Because no GLP-1 drug is approved for prediabetes, a request based on prediabetes alone may be denied. See our insurance and prior-authorization guide.

Paying yourself. Novo Nordisk lists Wegovy pens at $349 a month for self-pay after a $199 offer for the first two months. Lilly lists the Zepbound KwikPen at $299 to $699 a month depending on dose, with an offer that lowers some higher doses to as little as $449. Metformin is a generic drug and costs far less. Over several years of use, that gap adds up, which is part of why the Mayo lead author called metformin "safe, inexpensive" and still a first-line choice.

If you have prediabetes and are wondering about a GLP-1 drug, these questions can help focus the visit:

• How high is my risk? Is my A1C closer to 5.7% or 6.4%, and is my fasting glucose above 110 mg/dL?
• Do I meet the FDA criteria for Wegovy, Zepbound, or Foundayo based on my weight and other health problems?
• Would a Diabetes Prevention Program or metformin be a reasonable first step for me?
• If I start a GLP-1 drug, how long would I expect to stay on it, and what is the plan if I stop?
• Does my insurance cover it, and do I qualify for the Medicare GLP-1 Bridge?
• How often should my A1C be checked while I am on it?

For people with obesity and prediabetes, the trials show GLP-1 drugs kept most people's blood sugar out of the diabetes range while they were taking them. What is not known yet is whether years of preventive use lead to fewer heart attacks, kidney problems, or deaths, and how to keep the benefit after stopping.