# GLP-1 Drugs and Type 1 Diabetes: The Label Says Unevaluated. Three Trials Disagree.

By Ruth Calder, Enforcement Editor (https://glpwatchdog.org/authors/ruth-calder) | Published: 2026-08-19 | Updated: 2026-11-10

Canonical page: https://glpwatchdog.org/research/glp1-and-type-1-diabetes

> 36% of people with type 1 diabetes and obesity hit a composite of time-in-range, low-glucose safety and 5% weight loss, against 0% on placebo. Total daily insulin fell about a third in a second trial.

## Key points

- Semaglutide, 72 adults, 26 weeks: 36% hit the composite target vs 0% on placebo.
- Tirzepatide, 24 adults, 12 weeks: total daily insulin down 24.2 units, about 35% below placebo.
- Severe hypoglycemia was equal (2 events in each arm), and no diabetic ketoacidosis was reported.
- DKA is the risk that matters most, and about 120 people across three small trials cannot answer it.

The Wegovy label states that use in type 1 diabetes has not been evaluated.[5] Randomized trials have now evaluated it. In the largest recent one, 36% of people with type 1 diabetes and obesity hit a composite target of time-in-range, low-glucose safety and 5% weight loss, against **0%** on placebo.[1] In a smaller tirzepatide trial, total daily insulin fell by about a third.[2] The label’s statement and the trials’ results are both true, and understanding why is most of what you need here.

## Why the label and the literature disagree

A label is not a summary of what is known. It describes what a manufacturer submitted to a regulator and what the regulator approved on that basis. “Has not been evaluated” means no submission covered it, not that no one has studied it.

So prescribing here is off-label, which is legal, common and entirely a clinician’s judgment. It also means the usual scaffolding is missing: no approved dose for this population, no labeled guidance on adjusting insulin alongside it, and no post-marketing surveillance framed around type 1. The Wegovy sentence is also not universal: the current Mounjaro and Zepbound labels do not mention type 1 at all.[8][9]

## What the two trials found

*The two placebo-controlled parallel-group trials, both in adults with obesity and type 1 diabetes.*

|   | Semaglutide[1] | Tirzepatide[2] |
| --- | --- | --- |
| Design | 26 weeks, double-blind, 72 adults on automated insulin delivery | 12 weeks, phase 2, 24 adults (22 completed) |
| Dose | Up to 1 mg weekly | 2.5 mg then 5.0 mg weekly |
| Weight | −8.8 kg against placebo | −8.7 kg against placebo (8.8%) |
| HbA1c | −0.3 percentage points | −0.4 percentage points (P = 0.05) |
| Time in range 70–180 | +8.8 percentage points | — |
| Insulin dose | — | −24.2 units/day, about 35% below placebo |

The semaglutide trial’s primary endpoint deserves attention because of how it was built. To count as a success a participant had to achieve *all three* of: more than 70% of time in range, less than 4% of time below 70 mg/dl, and at least 5% weight loss. That is a demanding composite, deliberately including a safety condition rather than only benefits. 36% cleared it. Nobody on placebo did.[1]

## The safety question that matters most

In type 1 diabetes the fear is not weight or glucose control. It is **diabetic ketoacidosis**, the complication that ended enthusiasm for SGLT2 inhibitors in this population after they were shown to raise it. Anything that reduces insulin requirements in someone who makes no insulin invites that question.

> No diabetic ketoacidosis was reported in either trial. Neither was remotely large enough to rule it out.

A third, smaller trial adds a detail the other two did not. In a double-blind crossover study of 28 adults on automated insulin delivery, semaglutide titrated to 1 mg raised time in range by 4.8 percentage points without increasing time spent low. There was no ketoacidosis and no severe hypoglycemia. During semaglutide use, two recurrent episodes of euglycemic ketosis without acidosis were recorded.[6] Ketones rising while glucose looks normal is the early stage of the problem this population fears, and a glucose reading will not flag it.

Severe hypoglycemia was also unchanged: two events in the semaglutide group and two in the placebo group.[1] That is reassuring on the concern most people raise first, and it comes from 72 people over 26 weeks. Rare events need thousands of participant-years, and this literature has a few dozen.

> **A third less insulin is a large change to manage.** The tirzepatide trial recorded a 24.2 unit/day fall in total daily dose.[2] Getting that adjustment wrong in either direction is dangerous in type 1, and it is not a thing to do by feel. If a GLP-1 is being added to type 1 diabetes, the insulin plan should be adjusted by the team that manages it, not left to the prescriber who wrote the weight-loss script. There should be more glucose monitoring than usual through the transition, which means a prescription CGM with alarms; the [over-the-counter sensors](https://glpwatchdog.org/research/cgm-for-weight-loss-without-diabetes) are not cleared for insulin users.

## Who these trials were

Both enrolled people with type 1 diabetes *and* obesity (BMI 30 or above), a combination that is now common and was not always. The semaglutide trial required participants to be using an automated insulin delivery system, meaning a pump adjusting doses algorithmically.

That last detail limits how far the result travels. Automated delivery absorbs a great deal of variability, and a system that adjusts insulin every few minutes is a different safety context from injections on a fixed schedule. The reassuring hypoglycemia numbers were produced with that safety net in place.

## In young people, the weight moved and the glucose did not

A single-center retrospective study followed 24 adolescents and young adults aged 10 to 20 with type 1 diabetes and obesity who were prescribed a GLP-1 between 2019 and 2024.[4] What happened splits cleanly in two.

*Twelve months of GLP-1 treatment in 24 young people with type 1 diabetes.[4]*

| Measure | Change | P |
| --- | --- | --- |
| Body weight | −9.49 kg | < 0.0001 |
| BMI | −3.69 kg/m² | < 0.0001 |
| BMI Z-score | −0.30 | 0.04 |
| **Total daily insulin (pump users)** | **−21.42%** | 0.002 |
| Continuous monitor time in range | +7.96% | **0.08** |
| Time above range (180–250 mg/dL) | −3.04% | **0.06** |

> **Weight and insulin requirement moved decisively. The glucose control measures did not reach significance.** Time in range improved by nearly eight percentage points at P = 0.08, and time above range fell at P = 0.06. Both are close and neither crosses the line, in 24 people. In short, these young people needed substantially less insulin and weighed considerably less, and whether their glucose control improved is unresolved.

Twenty-four patients at one center, retrospectively, across five different drugs, is a small and heterogeneous basis for anything, and it is roughly what exists for this population. Three quarters were girls and young women, two thirds had public insurance, and most were using continuous monitors and pumps, which describes a group already engaged with intensive management. Young people with type 2 diabetes, by contrast, have [a randomized trial of tirzepatide in 10- to 17-year-olds](https://glpwatchdog.org/research/tirzepatide-in-ten-to-seventeen-year-olds).

## What a large real-world cohort adds

The trials above measured glucose control and insulin dose. A 2026 propensity-matched analysis of **4,088 people per group** with type 1 diabetes looked at harder outcomes, and found GLP-1-based therapy associated with lower all-cause mortality (HR 0.67), heart failure (HR 0.38), a composite cardiovascular endpoint (HR 0.61) and all-cause hospitalization (HR 0.70).[3]

Two safety results matter more here than the efficacy ones. **Diabetic ketoacidosis was not increased**, and DKA is the single largest fear about using these drugs in type 1 diabetes. And hypoglycemia risk was *lower*, at a hazard ratio of 0.72. Fewer than ten pancreatitis events occurred in either group.[3]

> It is observational and propensity-matched rather than randomized, and no significant differences appeared for myocardial infarction, ischemic stroke or chronic kidney disease. The authors call for randomized trials to confirm. What it does is put a large, matched dataset behind the safety question that has kept this class out of type 1 diabetes, and on DKA specifically the answer is reassuring.

## Where this leaves things

- **The evidence is real, small and short.** 72 people for 26 weeks and 24 for 12 weeks is a beginning, not a basis for routine use.
- **The weight and insulin effects are substantial** and are what would make this attractive to someone managing both conditions.
- **The DKA question is not closed.** The large cohort found no increase,[3] but it is observational. In the trials, absence in about 120 participants is weak evidence about a rare event, and one trial recorded ketosis without acidosis.[6]
- **The glucose benefit may not outlast the drug.** In a 12-week follow-up after the semaglutide trial, the 22 people who stopped lost time in range (median −3.6 points against +1.5 on placebo) and their glucose became more variable.[7]
- **Nothing here is approved**, so cost and coverage will be harder. Our [insurance section](https://glpwatchdog.org/insurance) covers what plans publish about off-label use.

Our [low blood sugar article](https://glpwatchdog.org/research/glp1-low-blood-sugar-risk) covers the hypoglycemia mechanics in more detail, including why these drugs are poor causes of it on their own and what changes when insulin is in the picture.

## Frequently asked questions

### Can you take a GLP-1 with type 1 diabetes?

It happens, off-label, and randomized trials have now studied it. The Wegovy label still says use in type 1 diabetes has not been evaluated, while type 1 goes unmentioned on the Mounjaro and Zepbound labels. That describes what the manufacturer submitted rather than what researchers have done. There is no approved dose or labeled insulin-adjustment guidance for this population.

### Does it work in type 1?

On the measures tested, yes. In a 26-week trial of 72 adults, 36% achieved a composite of over 70% time in range, under 4% time below 70 mg/dl, and at least 5% weight loss, against 0% on placebo. A 12-week tirzepatide trial found 8.7 kg more weight loss than placebo and cut total daily insulin by about 35%.

### What about diabetic ketoacidosis?

That is the right question, and the trials cannot answer it. No DKA was reported in the three randomized trials, which enrolled 72, 24 and 28 people for weeks rather than years, and one of them recorded two episodes of ketosis without acidosis on semaglutide. DKA is the complication that ended enthusiasm for SGLT2 inhibitors in type 1, and absence in about 120 people is weak evidence about a rare event. A large observational cohort of 4,088 people per group found DKA was not increased, which is reassuring but not randomized.

### Did it cause more low blood sugar?

Not in the trial that reported it: two severe hypoglycemia events in the semaglutide group and two in placebo. Every participant in that trial was using an automated insulin delivery system, which absorbs a lot of variability, so the result was produced with a safety net in place.

### Who should manage the insulin adjustment?

The team that manages the type 1 diabetes. A 35% fall in total daily insulin is a large change, getting it wrong in either direction is dangerous, and it should not be left to whoever wrote a weight-loss prescription.

## References

1. Shah VN, Akturk HK, Kruger D, et al. Semaglutide in Adults with Type 1 Diabetes and Obesity. NEJM Evidence. 2025. PMID: 40550013. https://pubmed.ncbi.nlm.nih.gov/40550013/
2. Snaith JR, Frampton R, Samocha-Bonet D, et al. Tirzepatide in Adults With Type 1 Diabetes: A Phase 2 Randomized Placebo-Controlled Clinical Trial. Diabetes Care. 2026. PMID: 41264593. https://pubmed.ncbi.nlm.nih.gov/41264593/
3. Tentolouris A, Filippatos C, Tepetes NI, et al. GLP-1 Receptor Agonist Therapy and Cardiorenal Outcomes in Type 1 Diabetes: A Propensity-Matched Real-World Analysis. Diabetes, Obesity and Metabolism. 2026. PMID: 42092242. https://pubmed.ncbi.nlm.nih.gov/42092242/
4. Gonzalez F, Reid MW, Garcia JF, et al. GLP-1 receptor agonists reduce body mass index and total daily insulin dose in youth with type 1 diabetes: a retrospective cohort study. Journal of Pediatric Endocrinology & Metabolism. 2026. PMID: 41353583. https://pubmed.ncbi.nlm.nih.gov/41353583/
5. Novo Nordisk Pharmaceutical Industries, LP WEGOVY (semaglutide) injection and tablets — prescribing information, SPL version 19. DailyMed, U.S. National Library of Medicine. 2026. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=ee06186f-2aa3-4990-a760-757579d8f77b
6. Pasqua MR, Tsoukas MA, Kobayati A, et al. Subcutaneous weekly semaglutide with automated insulin delivery in type 1 diabetes: a double-blind, randomized, crossover trial. Nature Medicine. 2025. PMID: 39794615. https://pubmed.ncbi.nlm.nih.gov/39794615/
7. 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, Obesity and Metabolism. 2026. PMID: 42675550. https://pubmed.ncbi.nlm.nih.gov/42675550/
8. Eli Lilly and Company MOUNJARO (tirzepatide) injection. Prescribing information. DailyMed, US National Library of Medicine. 2026. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=d2d7da5d-ad07-4228-955f-cf7e355c8cc0
9. Eli Lilly and Company ZEPBOUND (tirzepatide) injection. Prescribing information. DailyMed, US National Library of Medicine. 2026. https://dailymed.nlm.nih.gov/dailymed/drugInfo.cfm?setid=487cd7e7-434c-4925-99fa-aa80b1cc776b

---

Cite as: GLP Watchdog, Ruth Calder, "GLP-1 Drugs and Type 1 Diabetes: The Label Says Unevaluated. Three Trials Disagree.", published 2026-08-19, updated 2026-11-10. https://glpwatchdog.org/research/glp1-and-type-1-diabetes

Canonical HTML page: https://glpwatchdog.org/research/glp1-and-type-1-diabetes

Medical disclaimer: Nothing here is medical advice. We report on this subject; we do not treat anyone. Whether to begin a medication, supplement, or treatment, whether to adjust it, and whether to stop are questions for a clinician licensed to answer them, who knows your history.
