GLP-1 Medicines and Prediabetes: What the Trials Show About Getting Blood Sugar Back to Normal

Prediabetes is the stage where blood sugar has already drifted above normal but has not yet crossed into type 2 diabetes. Most people who have it never find out.

The CDC puts the count at 115.2 million Americans, and estimates that 8 in 10 adults with prediabetes do not know they have it. Prediabetes rarely causes symptoms. It shows up on a routine A1C or fasting glucose test, often as a footnote the patient is told to "keep an eye on."

GLP-1 medicines were not developed for prediabetes, but the big obesity trials enrolled thousands of people who had it, and researchers tracked what happened to their blood sugar. The answers are now in for semaglutide, tirzepatide and liraglutide. This article goes through those results, how they compare with the older lifestyle and metformin data, what happens after treatment stops, and where low-dose protocols fit, which is mostly in the gap where no trial has looked.

What blood sugar numbers count as prediabetes?

Three tests, three sets of cut-offs

Prediabetes is diagnosed when any one of three standard tests lands between the normal range and the diabetes range. The National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) lists the cut-offs below. A1C reflects average blood sugar over the past few months, fasting plasma glucose is a single morning snapshot, and the oral glucose tolerance test measures how well the body clears a sugar drink after 2 hours.

Source: NIDDK, Diabetes Tests & Diagnosis.

Can semaglutide bring prediabetes back to normal blood sugar?

The pooled STEP analysis

Semaglutide 2.4 mg returned most trial participants with prediabetes to normal blood sugar while they stayed on treatment. Leigh Perreault and colleagues pooled 3,375 adults with overweight or obesity from the STEP 1, 3 and 4 trials and published the glucose results in Diabetes Care in 2022. Of those, 1,536 started with prediabetes.

At week 68, normoglycemia was reached by 84.1% of participants with prediabetes on semaglutide in STEP 1 versus 47.8% on placebo. The figures were 89.5% versus 55.0% in STEP 3, and 89.8% versus 70.4% in STEP 4. STEP 4 is the odd one out because every participant took semaglutide during a 20-week run-in before being randomized, which explains its high placebo number. All three placebo groups still received diet and activity counselling, so the placebo bars show what that counselling achieved alone.

Does a GLP-1 stop normal blood sugar drifting into prediabetes?

The other direction

Fewer people with normal blood sugar slid into prediabetes on semaglutide than on placebo. The same Perreault analysis looked at participants who started with normoglycemia. By week 68, prediabetes had appeared in 2.9% of the semaglutide group versus 10.9% on placebo in STEP 1, 3.2% versus 5.8% in STEP 3, and 1.1% versus 5.0% in STEP 4. The absolute numbers are small because most people with normal glucose stay normal over 68 weeks, but the direction matched the prediabetes results.

How much did tirzepatide lower the risk of type 2 diabetes?

Three years of SURMOUNT-1

Tirzepatide cut new type 2 diabetes diagnoses in people with prediabetes from 13.3% on placebo to 1.3% over 176 weeks, a hazard ratio of 0.07. Ania Jastreboff and colleagues reported the 3-year results in the New England Journal of Medicine in 2025. SURMOUNT-1 enrolled 2539 people with obesity, and 1032 of them also had prediabetes. Those 1032 stayed on tirzepatide 5 mg, 10 mg or 15 mg weekly, or placebo, for 176 weeks.

The weight results explain much of the effect. Mean body weight fell by 12.3% on 5 mg, 18.7% on 10 mg and 19.7% on 15 mg at week 176, against 1.3% on placebo. Gastrointestinal side effects were the most common adverse events, mostly mild to moderate, and clustered in the first 20 weeks while doses were being stepped up.

What happens to blood sugar after a GLP-1 is stopped?

Part of the benefit fades within months

Some of the protection fades once treatment ends. SURMOUNT-1 included a 17-week off-treatment period after week 176. By the end of that period, the share of tirzepatide participants with type 2 diabetes had risen from 1.3% to 2.4%, while the placebo group sat at 13.7%. The hazard ratio moved from 0.07 to 0.12.

The off-treatment rate is still a fraction of placebo, which points to a real effect of the weight lost and the years spent with lower glucose. The rise within only 17 weeks also says that the drug is doing ongoing work. The STEP normoglycemia numbers carry the same caveat: Perreault's team measured them on treatment, before washout, so they describe what happens while people keep taking semaglutide.

What did the older liraglutide trial show?

SCALE Obesity and Prediabetes

Liraglutide 3.0 mg, a daily injection with smaller weight effects than semaglutide or tirzepatide, also lowered diabetes risk in prediabetes. The SCALE Obesity and Prediabetes trial, published by Carel le Roux and colleagues in The Lancet in 2017, randomized 2254 adults with prediabetes 2:1 to liraglutide or placebo alongside a reduced-calorie diet and more activity.

By week 160, 2% of the liraglutide group had been diagnosed with diabetes on treatment, against 6% on placebo, a hazard ratio of 0.21. Weight fell by 6.1% versus 1.9%. The trial has a large weakness: only about 50% of participants completed it, and those who withdrew were not followed afterward, so the true diabetes rate in the full group is unknown.

How do GLP-1 results compare with lifestyle change and metformin?

The Diabetes Prevention Program benchmark

No trial has put a GLP-1 head to head with a structured lifestyle program, but the Diabetes Prevention Program (DPP) remains the benchmark for prediabetes. Published by William Knowler and colleagues in the New England Journal of Medicine in 2002, the DPP randomized 3234 people with elevated glucose to placebo, metformin 850 mg twice daily, or a lifestyle program with goals of at least 7 percent weight loss and 150 minutes of activity a week.

Over an average 2.8 years, diabetes developed at 11.0 cases per 100 person-years on placebo, 7.8 on metformin and 4.8 with the lifestyle program. That is a 58 percent reduction for lifestyle and 31 percent for metformin. The DPP numbers use a different measure from the GLP-1 trials, so the bars below should not be set side by side with the SURMOUNT-1 percentages. The useful lesson is that a modest, sustained weight loss target already made a large difference.

Does a microdose GLP-1 protocol help with prediabetes?

Untested, and worth saying plainly

No trial has tested a microdose GLP-1 protocol for prediabetes, so its effect on blood sugar in this group is unknown. Every result in this article comes from full labelled regimens: semaglutide 2.4 mg weekly, tirzepatide 5 mg to 15 mg weekly, and liraglutide 3.0 mg daily.

The closest data point is the lowest tirzepatide dose. In the SURMOUNT-1 prediabetes group, 5 mg still produced a 12.3% average weight loss over 176 weeks. The published abstract reports diabetes onset for all tirzepatide doses combined, so it does not show whether 5 mg protected glucose as well as 15 mg. A microdose protocol typically sits below 5 mg, which is outside anything these trials measured.

The DPP offers an indirect clue. A lifestyle program built around a 7 percent weight loss goal cut diabetes incidence by 58 percent without any drug. If a lower-dose approach helps someone reach and hold modest weight loss with fewer side effects, that is a reasonable thing to discuss with a clinician. Claiming that a microdose itself prevents diabetes goes beyond the evidence, and anyone with prediabetes should have their A1C or fasting glucose rechecked on a schedule rather than assume a protocol is working.

What should someone with prediabetes track?

The markers the trials actually used

Frequently Asked Questions