GLP-1s for Type 1 Diabetes: What the Research Shows in 2026

If you live with type 1 diabetes, you have probably noticed that almost everything written about GLP-1 drugs assumes type 2. The trials, the labels, the marketing and most of the coverage are built around a disease where the pancreas still makes insulin.

Type 1 is different, and so is the evidence. This article covers what is actually known, what is genuinely unresolved, and what has changed since 2025.

This is general health information, not medical advice. Type 1 diabetes management, and any decision about adding a medicine to insulin, belongs with your diabetes care team.

First, the regulatory reality

No GLP-1 receptor agonist and no dual GIP/GLP-1 receptor agonist is approved for type 1 diabetes in the United States.

The Wegovy label puts it directly: “The use of WEGOVY in patients with type 1 diabetes mellitus or in combination with insulin has not been evaluated” [1].

Ozempic is indicated for three things, all in type 2 diabetes: improving glycemic control, reducing major adverse cardiovascular events in people with established cardiovascular disease, and reducing the risk of sustained eGFR decline, end-stage kidney disease and cardiovascular death in people with chronic kidney disease [2]. Mounjaro is indicated for glycemic control in type 2 diabetes in adults and children 10 and older, plus cardiovascular risk reduction in high-risk adults with type 2 diabetes [3]. Zepbound and Wegovy are obesity products.

Eli Lilly’s adolescent tirzepatide obesity trial SURMOUNT-ADOLESCENTS lists “Have Type 1 Diabetes” among its exclusion criteria [4].

So all of this is off-label. That does not mean unstudied. It means no regulator has reviewed the benefit-risk balance for this use, and if something goes wrong, there is no labeling to fall back on.

The reason clinicians are interested anyway

People with type 1 diabetes are not exempt from obesity. Roughly the same forces that drive weight gain in the general population operate here, plus insulin itself is an anabolic hormone. A person with type 1 diabetes and obesity faces the same cardiovascular and metabolic risks as anyone else, with insulin resistance layered on top of absolute insulin deficiency, sometimes called “double diabetes.”

The historical objection to using this drug class in type 1 was diabetic ketoacidosis. Anything that reduces insulin requirements, suppresses appetite and slows gastric emptying could in theory push someone toward DKA, especially during illness or if the drug is stopped abruptly.

That objection was reasonable. It has never been tested in a trial powered to answer it.

ADJUST-T1D: the first randomized evidence

ADJUST-T1D (NCT05537233) was a 26-week double-blind, multicenter, randomized, placebo-controlled trial in 72 adults with type 1 diabetes and a BMI of 30 or higher who were using automated insulin delivery systems [5]. It was funded by Breakthrough T1D and published in NEJM Evidence in August 2025 (online June 23, 2025). One detail gets lost in most coverage: the dose tested was semaglutide up to 1 mg weekly, the diabetes dose, not the 2.4 mg obesity dose [5].

The primary endpoint was a composite: more than 70 percent time in range, less than 4 percent time below range, and at least 5 percent weight loss [5].

Thirty-six percent of the semaglutide group achieved it. Zero percent of the placebo group did. The between-group difference was 36 percentage points (95 percent CI 20.6-52.2, P<0.001) [5]. HbA1c fell by 0.3 percentage points more on semaglutide (95 percent CI -0.6 to -0.05), time in range 70-180 mg/dL was 8.8 percentage points higher (3.9 to 13.7), and the difference in body weight was -8.8 kg, about 19.4 pounds (95 percent CI -10.6 to -7.0) [5]. That last figure is a between-group difference, not the raw amount every participant lost. There were two severe hypoglycemia events in each group, and no diabetic ketoacidosis was reported [5].

Two important limits. Seventy-two people is small. And every participant was using automated insulin delivery, which means the results do not transfer cleanly to someone on multiple daily injections or on a pump without automation [5].

A post hoc analysis published in Diabetes Care in May 2026 quantified the insulin side of this. Total daily insulin dose fell 22.6 percent over 26 weeks (95 percent CI -28.3 to -17.0), driven more by bolus insulin (-30.5 percent) than basal insulin (-15.6 percent). At week 4, 83 percent of the reduction was attributable to a direct drug effect and only 17 percent to weight loss; by week 26 the split was roughly even. Daily carbohydrate intake fell from 137 g to 107 g [13].

What happened when people stopped

A 12-week extension phase of ADJUST-T1D followed 16 participants who had been on semaglutide and discontinued it, and 22 who had been on placebo and took no GLP-1 drug during follow-up [6].

Compared with placebo, those who stopped semaglutide had [6]:

• A greater decline in time in range 70 to 180 mg/dL: median change -3.6 percent versus 1.5 percent (adjusted p=0.044)
• A greater increase in glucose standard deviation: 6.0 versus 1.2 mg/dL (adjusted p=0.044)
• A greater increase in coefficient of variation: 3.1 versus 1.1 percent (adjusted p=0.044)

Time above range increased numerically but was not significant after adjustment, and there were no significant differences in hypoglycemia metrics [6].

The authors concluded that “glycaemic benefits achieved during treatment may not be fully sustained following withdrawal” [6]. Sixteen people is a very small group, but the finding matches what happens in every other population: stopping undoes the effect.

What the national data show

Denmark: 879 people on semaglutide

Danish researchers used national registries covering 2018 to 2024 to identify everyone with type 1 diabetes who started semaglutide, matching them 1:4 to unexposed controls with type 1 diabetes [7].

They found 879 people who started semaglutide: 622 on multiple daily injections and 257 on an insulin pump [7].

Results [7]:

• HbA1c fell 5.7 mmol/mol, about 0.52 percentage points, during the first six months and stayed stable after that. Controls were unchanged.
• No increased rate of hospitalization for hypoglycemia: hazard ratio 0.64 (95 percent CI 0.35-1.19).
• No increased rate of hospitalization for diabetic ketoacidosis: hazard ratio 0.73 (0.34-1.57).
• The cumulative probability of still being on semaglutide at one year was 50 percent.
• The median dose redeemed was 1.0 mg, well below the obesity dose.
• There was no difference in HbA1c reduction between injection users and pump users (P=0.42).

That 50 percent one-year adherence figure is arguably the most striking number in the study. Half the people who started stopped within a year.

Sheffield: 142 people on tirzepatide

A UK retrospective cohort at Sheffield Teaching Hospitals compared 142 adults with type 1 diabetes and a BMI of 30 or higher started on tirzepatide against 50 age-, sex- and BMI-matched controls, followed for a median 365 days [8].

Mean age was 42.8, mean BMI 37.4, mean HbA1c 64.3 mmol/mol, and 32 percent were male. Most, 72.4 percent, were on 5 mg weekly [8].

Compared with controls, the tirzepatide group achieved [8]:

• HbA1c reduction of 6.5 mmol/mol, about 0.6 percentage points (p<0.001)
• Weight reduction of 13.4 kg, about 29.5 pounds (p<0.001)
• Total daily insulin dose reduction of 29.9 units (p<0.001)
• Systolic blood pressure down 12.6 mmHg, diastolic down 4.1 mmHg
• Total cholesterol down 0.8 mmol/L, non-HDL cholesterol down 0.6 mmol/L, triglycerides down 0.7 mmol/L

Side effects were common but 89.4 percent were still on tirzepatide at one year, and there was no between-group difference in severe hypoglycemia, ketoacidosis or hospitalization rates [8].

The contrast in persistence between the Danish semaglutide cohort (50 percent at one year) and the Sheffield tirzepatide cohort (89.4 percent) is large. It probably reflects differences in how the drug was prescribed, supported and dosed rather than a property of the drugs themselves.

The three open safety questions

Diabetic ketoacidosis

Neither the Danish semaglutide cohort nor the Sheffield tirzepatide cohort found an increased rate [7][8]. Neither did ADJUST-T1D [5].

But all three are either small or observational, and hospitalization is a crude way to measure DKA, since it misses episodes managed at home or in urgent care. The insulin dose reductions seen in these studies are large, which is precisely the situation in which DKA risk rises if a dose is missed, a pump site fails, or illness intervenes.

TTT1, Triple Therapy for Type 1 Diabetes, is an international phase 3 trial testing insulin plus semaglutide plus dapagliflozin in adults with type 1 diabetes; its design and methods were published in Diabetes, Obesity and Metabolism in 2026 [9]. Adding an SGLT2 inhibitor raises the DKA stakes further, which is why the trial’s safety framework matters.

Retinopathy

Both the Ozempic and Wegovy labels carry a Warnings and Precautions item on diabetic retinopathy complications in patients with type 2 diabetes, advising that people with a history of diabetic retinopathy be monitored [1][2]. The mechanism, rapid improvement in glucose control, applies to type 1 as well.

A 2026 report in Diabetes Technology and Therapeutics examined incident diabetic retinopathy after one year of adjunctive tirzepatide in adults with type 1 diabetes [10]. The labels do not tell patients what to do here; they instruct prescribers to monitor patients with a history of diabetic retinopathy for progression [1][2].

Body composition and bone

A 2026 study in Diabetes Metabolism Research and Reviews examined weight-loss-dependent changes in body composition and bone health in people with obesity and type 1 diabetes treated with liraglutide, semaglutide or tirzepatide [11]. People with type 1 diabetes already have elevated fracture risk relative to the general population, so this is a more pointed question here than elsewhere.

What about adolescents with type 1 diabetes?

Very little published data. A 2026 case report in Pediatrics described the effects of semaglutide on weight and insulin requirements in two adolescents with type 1 diabetes [12]. Two patients.

Note that the tirzepatide adolescent obesity trials explicitly exclude type 1 diabetes [4], so that gap is not closing soon.

If you are considering this

These are reasonable things to raise with your diabetes team, based on what the studies actually describe:

• Whether the goal is weight, glucose variability, insulin dose or cardiovascular risk, because the evidence is strongest for weight and time in range.
• How insulin adjustment will be handled, since the dose reductions in these studies were large and rapid [5][8].
• What the sick-day and pump-failure plan looks like, given the DKA question.
• Whether you have retinopathy, and what eye monitoring would look like [1][2].
• What happens if you need to stop, and how glucose control would be expected to shift [6].
• Whether your device setup matches the trial evidence: the randomized evidence is in people on automated insulin delivery [5].

On insulin dosing, the labels are explicit about whose job it is. The Wegovy and Ozempic labels instruct prescribers to monitor blood glucose before and during treatment and to consider reducing the dose of concomitant insulin or an insulin secretagogue to reduce hypoglycemia risk [1][2]. That instruction is written to the prescriber, not to the patient.

Sources

1. WEGOVY (semaglutide) US Prescribing Information. Novo Nordisk. https://www.novo-pi.com/wegovy.pdf
2. OZEMPIC (semaglutide) US Prescribing Information. Novo Nordisk. https://www.novo-pi.com/ozempic.pdf
3. MOUNJARO (tirzepatide) US Prescribing Information. Eli Lilly. https://uspl.lilly.com/mounjaro/mounjaro.html
4. SURMOUNT-ADOLESCENTS, NCT06075667, key exclusion criteria. ClinicalTrials.gov. https://clinicaltrials.gov/study/NCT06075667
5. ADJUST-T1D. NEJM Evidence, 2025; NCT05537233. https://evidence.nejm.org/doi/full/10.1056/EVIDoa2500173 and https://clinicaltrials.gov/study/NCT05537233
6. Montaser E, Bhasin K, Kruger D, et al. Glycaemic Changes After Semaglutide Discontinuation in Adults With Type 1 Diabetes Using Automated Insulin Delivery: Analysis of the ADJUST-T1D Extension Study. Diabetes Obes Metab. 2026 Aug 31. https://doi.org/10.1111/dom.71287
7. Würtz Yazdanfard PD, Kosjerina V, Wood-Kurland HK, et al. Effectiveness and Safety of Semaglutide in Type 1 Diabetes: A Danish Nationwide Cohort Study (2018-2024). Lancet Reg Health Eur. 2026 May 18. https://doi.org/10.1016/j.lanepe.2026.101716
8. Berry SA, Goodman IR, McNally E, Elliott J, Iqbal A. Real-World Effectiveness and Safety of Tirzepatide in Type 1 Diabetes and Obesity. Diabetes Obes Metab. 2026 Oct. https://doi.org/10.1111/dom.71080
9. Timmons JG, Ghanim H, Boyle JG, et al. Insulin, Semaglutide and Dapagliflozin in Adults With Type 1 Diabetes: Design and Methods of Triple Therapy for Type 1 Diabetes (TTT1). Diabetes Obes Metab. 2026 Oct. https://doi.org/10.1111/dom.71076
10. Polsky S, Beck E, Shah A, Cengiz D, Garg SK. Incident Diabetic Retinopathy after Adjunctive Tirzepatide Treatment for 1 Year in Adults with Type 1 Diabetes. Diabetes Technol Ther. 2026 Jul 14. https://doi.org/10.1177/15209156261468108
11. Al Ozairi E, Irshad M, Alkandri J, et al. Weight Loss-Dependent Changes in Body Composition and Bone Health in People With Obesity and Type 1 Diabetes Treated With Liraglutide, Semaglutide, or Tirzepatide. Diabetes Metab Res Rev. 2026 Sep. https://doi.org/10.1002/dmrr.70213
12. Schwarz CR, Vilsbøll T, Aistrup MO, Kloppenborg JT. Effects of Semaglutide on Weight and Insulin Requirements in Two Adolescents With Type 1 Diabetes. Pediatrics. 2026 Jun 1. https://doi.org/10.1542/peds.2025-074286
13. Karakus KE, Akturk HK, Kruger D, et al. Effect of Semaglutide on Insulin Dose Reduction in Adults With Type 1 Diabetes and Obesity Using Automated Insulin Delivery Systems: ADJUST-T1D Post Hoc Analysis. Diabetes Care. 2026 May 1;49(5):718-723. https://doi.org/10.2337/dc25-2249
14. SURMOUNT-ADOLESCENTS eligibility criteria (exclusion: “Have Type 1 Diabetes”), NCT06075667. ClinicalTrials.gov. https://clinicaltrials.gov/study/NCT06075667