No GLP-1 drug is approved in the United States for prediabetes. What exists is trial evidence from people with obesity who also had prediabetes, collected while the drugs were tested for weight management. In the longest of those trials, SURMOUNT-1, 1.3% of people on tirzepatide and 13.3% on placebo were diagnosed with type 2 diabetes over 176 weeks. Liraglutide and semaglutide have their own results, each measured differently.
This page sets the three trials side by side and states what each could and could not show. Labels were read on DailyMed on 17 October 2026; trial figures are from the publications named below. It reports what the documents say; it does not advise on treatment.
What the labels say: nothing, by design
Nine current GLP-1 labels were searched for "prediabetes", "pre-diabetes", "impaired glucose" and "impaired fasting" on 17 October 2026. None has a prediabetes indication. The term appears twice: in the Saxenda label as a factor the trial statistics adjusted for, and in the Zepbound label's table of baseline characteristics. The main approved uses are type 2 diabetes (Ozempic, Mounjaro, Victoza, Trulicity, Xultophy) and weight management in adults meeting body-mass-index criteria (Wegovy, Zepbound, Foundayo, Saxenda). A person with prediabetes reaches these drugs, if at all, through the weight-management indication.
The three trials
| SURMOUNT-1 (tirzepatide) | SCALE Obesity and Prediabetes (liraglutide 3.0 mg) | STEP 1, 3 and 4 (semaglutide 2.4 mg) | |
|---|---|---|---|
| Who with prediabetes | 1,032 adults with obesity and prediabetes | 2,254 adults with obesity or overweight and prediabetes | 1,536 participants with prediabetes at baseline, pooled from three trials |
| Duration on treatment | 176 weeks, then 17 weeks off | 160 weeks | 68 weeks |
| Outcome measured | Diagnosis of type 2 diabetes | Diagnosis of diabetes while on treatment | Glucose back to normal (normoglycaemia) |
| Drug | 1.3% diagnosed | 2% (26 of 1,472) | 84.1% / 89.5% / 89.8% normal |
| Placebo | 13.3% diagnosed | 6% (46 of 738) | 47.8% / 55.0% / 70.4% normal |
| Hazard ratio | 0.07 (95% CI 0.0 to 0.1) | 0.21 (0.13 to 0.34) | — |
| Weight change, drug vs placebo | -12.3% to -19.7% by dose vs -1.3% | -6.1% vs -1.9% | not reported in this analysis |
| Sponsor | Eli Lilly | Novo Nordisk | Novo Nordisk |
Sources: Jastreboff et al., NEJM 2025 (SURMOUNT-1, NCT04184622); le Roux et al., Lancet 2017 (SCALE); Perreault et al., Diabetes Care 2022 (STEP). Read 2026-10-17.
Reading across the row is the mistake to avoid. The three trials measured different things. SURMOUNT-1 and SCALE counted new diagnoses of diabetes; the STEP analysis counted how many people's glucose returned to the normal range. The populations, the diagnostic rules and the durations differ, and no trial compared two of these drugs. The figures show each drug against its own placebo, nothing more.
The absolute numbers
A hazard ratio of 0.07 sounds total, so the absolute figures matter:
- SURMOUNT-1: 13.3% minus 1.3% is 12.0 percentage points over 176 weeks — roughly one diagnosis avoided for every 8 people treated for about three and a half years.
- SCALE: on the on-treatment counts, 6.2% minus 1.8% is 4.4 points over 160 weeks, roughly one in 23.
Both are our arithmetic from the published counts, not figures the papers state. The gap between them partly reflects the placebo groups: 13.3% of SURMOUNT-1's placebo group developed diabetes against about 6% of SCALE's.
What happened after stopping
SURMOUNT-1 is the only one of the three that measured a period off treatment. Seventeen weeks after stopping, 2.4% of the former tirzepatide group and 13.7% of the placebo group had type 2 diabetes (hazard ratio 0.12). The gap was still wide at that point, but 17 weeks is short, and the paper reports nothing after it. The broader record on what happens when these drugs stop — weight regain in particular — is on stopping a GLP-1 drug.
The SCALE caveat
SCALE lost many participants: 47% of the liraglutide group and 55% of the placebo group withdrew before week 160, and, as its authors state, withdrawn individuals were not followed up after they stopped. Its diabetes counts are therefore counts among people still on study drug. That is the trial's own stated limitation, and it means the true difference in the whole randomised population is not known.
The comparison people ask about: metformin
Metformin is the drug most often discussed for prediabetes, and its evidence is older and from a different design. In the Diabetes Prevention Program (Knowler et al., NEJM 2002), diabetes incidence was 11.0 cases per 100 person-years on placebo, 7.8 on metformin and 4.8 with an intensive lifestyle programme — reductions of 31% and 58%. No trial has put a GLP-1 drug against metformin or against a structured lifestyle programme in prediabetes, so those numbers cannot be ranked against the table above.
What is not established
- A prediabetes indication for any GLP-1 drug in the US, and therefore any label-based use for people with prediabetes who do not meet weight-management criteria.
- Long-term prevention. No trial here reports beyond about three and a half years, or more than 17 weeks off treatment.
- Outcomes beyond diagnosis. None of the three was designed to show fewer heart attacks, kidney failure or deaths in people with prediabetes.
- People without obesity. All three enrolled people with obesity or overweight.
Whether a GLP-1 drug, metformin, a lifestyle programme or none of these suits a person with prediabetes is a decision for their clinician. How the drugs work is on how GLP-1 drugs work; the tirzepatide weight results trial by trial are on Zepbound weight-loss results; the liraglutide-versus-semaglutide trial is on semaglutide vs liraglutide; coverage rules are on GLP-1 insurance coverage.
Sources
- Jastreboff AM et al. "Tirzepatide for Obesity Treatment and Diabetes Prevention." N Engl J Med 2025 (published online 2024), SURMOUNT-1, NCT04184622, PubMed 39536238.
- le Roux CW 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." Lancet 2017, PubMed 28237263.
- Perreault L 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;45(10):2396-2405, PubMed 35724304.
- Knowler WC et al. "Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin." N Engl J Med 2002, PubMed 11832527.
- US prescribing information on DailyMed, read 2026-10-17: OZEMPIC, WEGOVY, MOUNJARO, ZEPBOUND, FOUNDAYO, SAXENDA, VICTOZA, TRULICITY, XULTOPHY 100/3.6 — searched for "prediabetes", "pre-diabetes", "impaired glucose" and "impaired fasting".