GLP-1 Microdosing and Prediabetes: What the Trials Show About Preventing Type 2 Diabetes

A borderline HbA1c is one of the most common reasons people first ask about GLP-1 medicines. The number sits in a gray zone: too high to ignore, too low for a diabetes diagnosis, and usually paired with advice to lose some weight and move more. The CDC's National Diabetes Statistics Report puts the number of US adults with prediabetes at 115.2 million.

This article looks at what GLP-1 microdosing and prediabetes have in common in the research: what the tirzepatide, semaglutide and liraglutide trials found in people with prediabetes, how those results compare with lifestyle change and metformin, what happened when people stopped, and how much of it applies to a low dose. The full-dose evidence is strong. The low-dose evidence mostly does not exist yet.

What is prediabetes, and why does it progress?

Prediabetes is blood sugar that runs above normal but below the threshold for type 2 diabetes. A common working definition, used in the SURMOUNT-5 analysis by Galindo and colleagues (J Endocrinol Invest, 2026), is a fasting glucose of 100–125 mg/dL or an HbA1c of 5.7–6.4%. Normal glucose, which researchers call normoglycemia, means an HbA1c below 5.7% together with a fasting glucose below 100 mg/dL.

Prediabetes usually reflects insulin resistance, often driven by excess body fat, with the pancreas working harder to keep glucose in range. Some people stay in that zone for years and some return to normal. Others progress to type 2 diabetes. The placebo arm of SURMOUNT-1 gives a sense of the pace: 13.3% of people with obesity and prediabetes who received placebo were diagnosed with type 2 diabetes within 176 weeks. For background on how GLP-1 medicines act on glucose and insulin, see our guide to GLP-1 microdosing and insulin sensitivity.

Can tirzepatide prevent type 2 diabetes in people with prediabetes?

In the longest trial so far, tirzepatide sharply cut new type 2 diabetes diagnoses compared with placebo over three years of treatment. Jastreboff and colleagues reported the 3-year results of SURMOUNT-1 in the New England Journal of Medicine (2025). Of the 2539 adults with obesity in the trial, 1032 also had prediabetes. They were randomized to weekly tirzepatide 5 mg, 10 mg or 15 mg, or placebo, for 176 weeks, then followed for a 17-week off-treatment period.

During treatment, 1.3% of the tirzepatide group were diagnosed with type 2 diabetes against 13.3% of the placebo group, a hazard ratio of 0.07. The trial's safety findings were familiar: the most common side effects were gastrointestinal, mostly mild to moderate, and they clustered in the dose-escalation period during the first 20 weeks. No new safety signals turned up over the three years.

Does the lowest tirzepatide dose still work for prediabetes?

Tirzepatide 5 mg, the lowest maintenance dose in the trial, still produced substantial weight loss in people with prediabetes, though less than the higher doses. At 176 weeks, average weight change was 12.3% down on 5 mg, 18.7% on 10 mg and 19.7% on 15 mg, compared with 1.3% on placebo.

The step from 5 mg to 10 mg added more weight loss than the step from 10 mg to 15 mg, which is the kind of curve that makes lower doses interesting. The catch is what the abstract does not report. Diabetes onset was published for all tirzepatide doses pooled, so the trial does not tell us how many people on 5 mg alone progressed. And 5 mg is still a full approved dose. Microdosing usually means staying below it, and no prediabetes trial has looked at that range. Our article on microdosing tirzepatide for maintenance covers what is known about stepping down.

How many people return to normal blood sugar on semaglutide?

Most of them. Perreault and colleagues (Diabetes Care, 2022) pooled 3,375 adults with overweight or obesity from STEP 1, 3 and 4. Among the 1,536 who started with prediabetes, semaglutide 2.4 mg weekly for 68 weeks returned far more people to normoglycemia than placebo.

The rates were 84.1% versus 47.8% in STEP 1, 89.5% versus 55.0% in STEP 3, and 89.8% versus 70.4% in STEP 4. The placebo numbers are high because every participant also got lifestyle counseling, and STEP 4 began with a 20-week semaglutide run-in for everyone. The drug also kept people out of prediabetes in the first place. Among those who started with normal glucose, only 2.9% developed prediabetes on semaglutide in STEP 1, versus 10.9% on placebo.

What did the liraglutide prediabetes trial show over three years?

Liraglutide also lowered progression to diabetes, by a smaller weight margin and with a big dropout problem. The SCALE Obesity and Prediabetes trial, reported by le Roux and colleagues in The Lancet (2017), randomized 2254 adults with prediabetes to liraglutide 3.0 mg daily or placebo, on top of diet and activity advice.

By week 160, 2% of the liraglutide group (26 of 1472) had been diagnosed with diabetes while on treatment, against 6% of the placebo group (46 of 738). The hazard ratio was 0.21, and the time to onset was 2.7 times longer with liraglutide. Weight change was 6.1% down on liraglutide and 1.9% down on placebo. The weakness is that 47% of the liraglutide group and 55% of the placebo group withdrew, and people who left were not followed up, so the true rates outside the trial are uncertain.

Is a GLP-1 better than lifestyle change or metformin for prediabetes?

Nobody has run that comparison directly, so the honest answer is unknown. The benchmark for prediabetes is still the Diabetes Prevention Program (Knowler and colleagues, NEJM, 2002), which randomized 3234 people with raised fasting and post-meal glucose to placebo, metformin 850 mg twice daily, or a lifestyle program aiming for at least 7% weight loss and 150 minutes of activity per week.

Lifestyle change cut diabetes incidence by 58 percent and metformin by 31 percent compared with placebo. The GLP-1 trials report their results differently, so the percentages cannot be lined up directly. What they share is a design: every GLP-1 arm also received diet and activity advice. The fair reading is that a GLP-1 adds to lifestyle change, and the biggest effects so far came with the largest weight losses.

Is tirzepatide or semaglutide better for prediabetes?

In the one head-to-head trial, more people with prediabetes reached normal glucose on tirzepatide. Galindo and colleagues ran a post hoc analysis of 425 people with prediabetes in SURMOUNT-5, which compared tirzepatide at its maximum tolerated dose (10 mg or 15 mg) with semaglutide at its maximum tolerated dose (1.7 mg or 2.4 mg). Average starting HbA1c was 5.86%.

At week 72, 89.9% on tirzepatide had reverted to normoglycemia against 76.2% on semaglutide. HbA1c fell 0.60% versus 0.48%, and weight fell 21.5% versus 14.5%. Both doses were at the top of their ranges, though, so this says nothing about how the two drugs compare at low doses. Post hoc analyses are also weaker evidence than a trial built to answer the question.

What happens to prediabetes after stopping a GLP-1?

Some of the protection fades within months. In SURMOUNT-1, the share of tirzepatide participants with type 2 diabetes rose from 1.3% to 2.4% after 17 weeks off the drug, while the placebo group stood at 13.7%. The hazard ratio moved from 0.07 to 0.12. Tirzepatide still looked far better than placebo, but the gap had started to close.

The STEP 1 extension shows where that drift leads. Wilding and colleagues (Diabetes Obes Metab, 2022) followed 327 people for a year after semaglutide and lifestyle support stopped at week 68. People on semaglutide had lost 17.3% of their weight and regained 11.6 percentage points of it by week 120, and the cardiometabolic improvements moved back toward baseline. For prediabetes, that suggests the benefit depends on staying on some form of treatment. Whether a lower maintenance dose can hold the gains is a question people ask often, and it has not yet been answered in a prediabetes trial.

What does this mean for GLP-1 microdosing and prediabetes?

The prevention data are real but were all gathered at approved doses, so any low-dose plan is an extrapolation. A few points follow from the trials:

• The lowest dose with prediabetes data is tirzepatide 5 mg weekly, and its diabetes-prevention result was not published separately from the higher doses.
• Weight loss tracks with the glucose benefit across these trials. A dose that produces little weight loss should not be expected to match the trial numbers.
• The drug was always given alongside diet and activity advice, and lifestyle change on its own cut diabetes incidence by 58 percent in the Diabetes Prevention Program.
• Stopping brings some risk back, so a plan should cover what happens after the first year, not just the start.
• Track the numbers that define prediabetes: HbA1c and fasting glucose, checked with a clinician, rather than the scale alone.

People with prediabetes who also get shaky or foggy a few hours after meals may be dealing with something different. Our article on reactive hypoglycemia covers that pattern.

Frequently asked questions

Clinical references

1. Jastreboff AM, le Roux CW, Stefanski A, et al. Tirzepatide for Obesity Treatment and Diabetes Prevention. New England Journal of Medicine. 2025. doi:10.1056/NEJMoa2410819
2. Perreault L, Davies M, Frias JP, et al. Changes in Glucose Metabolism and Glycemic Status With Once-Weekly Subcutaneous Semaglutide 2.4 mg Among Participants With Prediabetes in the STEP Program. Diabetes Care. 2022. doi:10.2337/dc21-1785
3. le Roux CW, Astrup A, Fujioka K, et al. 3 years of liraglutide versus placebo for type 2 diabetes risk reduction and weight management in individuals with prediabetes: a randomised, double-blind trial. The Lancet. 2017. doi:10.1016/S0140-6736(17)30069-7
4. Galindo RJ, Aronne LJ, Horn DB, et al. Reversion to normoglycemia with tirzepatide vs semaglutide in participants with obesity and prediabetes: a post hoc analysis of SURMOUNT-5. Journal of Endocrinological Investigation. 2026. doi:10.1007/s40618-026-02895-3
5. Knowler WC, Barrett-Connor E, Fowler SE, et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. New England Journal of Medicine. 2002. doi:10.1056/NEJMoa012512
6. Wilding JPH, Batterham RL, Davies M, et al. Weight regain and cardiometabolic effects after withdrawal of semaglutide: The STEP 1 trial extension. Diabetes, Obesity and Metabolism. 2022. doi:10.1111/dom.14725
7. Centers for Disease Control and Prevention. National Diabetes Statistics Report. cdc.gov