Semaglutide withdrawal in type 1 diabetes

A new extension analysis from the ADJUST T1D trial found that adults with type 1 diabetes and obesity who had previously received semaglutide had lower time in range and more variable glucose readings over the next 12 weeks after treatment ended than the placebo group. The result is a small, post trial comparison, not a rule for individual care. It does show why a change in a glucose lowering medicine belongs in a clinician led plan for people using automated insulin delivery^[1].

What the ADJUST T1D extension measured

ADJUST T1D was a randomized trial in adults with type 1 diabetes, obesity, and automated insulin delivery. The original 26 week trial evaluated semaglutide alongside that technology. Its extension followed continuous glucose monitor data after the trial period. Sixteen participants who had received semaglutide no longer received it, while 22 people previously assigned to placebo remained without a glucagon like peptide 1 receptor agonist during follow up^[1].

The study is relevant to a narrow group. It does not test semaglutide as an approved treatment for type 1 diabetes, and it does not establish what will happen for every person who stops it. The Food and Drug Administration labels semaglutide products for specified type 2 diabetes, cardiovascular risk, kidney disease, or weight management indications, depending on the product. Type 1 diabetes is not among those labelled indications^[2]. For a broader evidence overview, see our semaglutide guide.

What changed after treatment ended

From week 26 to the extension phase, the semaglutide group had a median 3.6 percentage point decline in time in range from 70 to 180 mg per dL. The placebo group had a median 1.5 percentage point increase. After adjustment for multiple comparisons, the reported difference had a p value of 0.044^[1]. Time in range is a continuous glucose monitor metric, not a complete account of diabetes control, but it is a useful way to describe how often readings sit within a specified target interval.

Glucose standard deviation rose by a median 6.0 mg per dL compared with 1.2 mg per dL in the placebo group. Coefficient of variation rose by 3.1 percent versus 1.1 percent. The paper reports that time above range increased numerically, but that comparison did not remain statistically significant after adjustment. No significant between group differences appeared in monitor based hypoglycaemia measures^[1].

How to read a small extension study

The direction of the finding is plausible, yet the evidence has clear limits. Only 38 people contributed to the extension comparison, and the analysis compared groups after a trial phase rather than randomly assigning people to continue or stop treatment at that point. The published abstract reports adjusted analyses and a sensitivity analysis, but it cannot separate every influence on glucose readings during those 12 weeks. Changes in food intake, illness, activity, insulin settings, or device use may matter in an individual record.

The main ADJUST T1D trial is useful context because it examined semaglutide during treatment rather than after it ended. In that 26 week randomized trial, investigators studied adults with type 1 diabetes, obesity, and automated insulin delivery. The extension paper describes its own results as evidence that glycaemic benefits achieved during treatment may not be fully sustained once treatment is withdrawn^[1]. That wording is appropriately cautious. It is an observed association in this study population, not proof of a universal rebound pattern.

Why automated insulin delivery does not remove the clinical question

Automated insulin delivery uses glucose data and an algorithm to adjust insulin delivery within its programmed limits. It can improve aspects of glucose management, but it does not make medication changes consequence free. In this study, participants were already using that technology, and the group difference in time in range still appeared in the follow up data. That makes the result more specific, not more generalizable to every form of diabetes care.

For people with type 1 diabetes, insulin remains essential. The National Institute of Diabetes and Digestive and Kidney Diseases explains that type 1 diabetes occurs when the body makes little or no insulin and that people with the condition need insulin every day to stay alive^[3]. Semaglutide does not replace insulin in this context. A care team that knows the person's glucose data, insulin needs, and device settings is the appropriate place to interpret a medication change. Our article on semaglutide and type 1 diabetes in a Danish cohort adds observational context but does not change that boundary.

What we do not yet know

This extension cannot tell readers how long a change in time in range or glucose variability might last, whether the findings apply beyond adults with obesity who use automated insulin delivery, or which individual factors account for a given pattern. It also cannot establish the safety or benefit of semaglutide for type 1 diabetes generally. The published sample was small, the follow up was 12 weeks, and the study was not designed as a new randomized withdrawal trial. More work is needed before these findings can support broad clinical conclusions.

What this result means in practice

The most defensible takeaway is narrow: in this ADJUST T1D extension, ending semaglutide was associated with a modest fall in time in range and higher glucose variability over 12 weeks compared with the placebo group. That is useful information for a clinical conversation, especially when automated insulin delivery is part of the picture. It is not a self management protocol. Anyone with type 1 diabetes considering a change to prescribed treatment should discuss it with their diabetes clinician rather than relying on a trial summary or an online account.

Related reading: semaglutide evidence overview, United States status, Denmark status, the regulation hub, the peptide library, and our Danish cohort analysis, semaglutide research, United States regulatory context, and Denmark regulatory context, semaglutide article, United States semaglutide page, and Denmark semaglutide page, and another semaglutide evidence entry.