Blood sugar gains fade after stopping semaglutide in type 1 diabetes extension study

A 12-week extension of the ADJUST-T1D trial found that adults with type 1 diabetes and obesity who stopped weekly semaglutide saw their continuous glucose monitoring numbers worsen, while participants who had taken placebo improved slightly. The analysis was published August 31, 2026, in Diabetes, Obesity and Metabolism ^[1][2].

The extension followed 16 people who had been randomly assigned to semaglutide in the original trial and stopped the drug, and 22 people who had been assigned to placebo and did not start a GLP-1 receptor agonist during follow-up ^[1][2]. Continuous glucose monitoring data were collected for 12 weeks after the trial's 26-week treatment period ^[1][2].

Compared with the placebo group, people who stopped semaglutide had a median change in time in range, the share of the day spent with glucose between 70 and 180 mg/dL, of -3.6 percentage points, versus +1.5 percentage points for the placebo group. The difference was statistically significant (adjusted p = 0.044) ^[1][2]. The semaglutide group also had larger increases in glucose standard deviation (6.0 vs. 1.2 mg/dL) and coefficient of variation (3.1% vs. 1.1%), both with an adjusted p = 0.044, which the authors describe as worsening glycemic variability ^[1][2]. Time above range, meaning glucose above 180 mg/dL, rose numerically but did not remain significant after adjustment (adjusted p = 0.064), and the researchers found no significant differences between the groups in CGM measures related to low blood sugar ^[1][2].

In the original ADJUST-T1D trial, 72 adults with type 1 diabetes and obesity using automated insulin delivery systems were randomly assigned in a 1:1 ratio to semaglutide or placebo. Semaglutide started at 0.25 mg once weekly for four weeks, then was raised to 0.5 mg and then 1 mg weekly through the 26-week treatment period, with temporary dose reductions allowed for intolerance ^[1]. The trial's primary composite outcome, which required time in range above 70%, time below range under 4%, and weight loss of at least 5%, was met by 36% of the semaglutide group and 0% of the placebo group (p < 0.001) ^[1]. A post hoc analysis found a 30% reduction in insulin requirements with semaglutide compared with placebo ^[1]. The trial studied semaglutide as an off-label add-on to automated insulin delivery ^[1].

Studies of other incretin drugs have shown comparable patterns after treatment stops. A four-week off-treatment follow-up of the SURPASS-1 trial found HbA1c rose about 0.2% after tirzepatide was stopped, and a systematic review and meta-analysis reported HbA1c increases of 0.25% in people with obesity and 0.65% in people with type 2 diabetes after GLP-1 receptor agonist discontinuation ^[1]. The STEP 1 extension found cardiometabolic improvements reverted toward baseline a year after semaglutide withdrawal ^[1].

The extension included 38 people in total, and the authors say their findings apply to adults with type 1 diabetes and obesity who used semaglutide in the context of weight management. They state the results may not be directly generalizable to other groups receiving semaglutide, including people using it off-label for insulin resistance or excess weight without obesity ^[1]. How long the decline continues, whether it levels off, and whether restarting semaglutide restores earlier control have not been reported from this extension. Changes in weight and insulin dose during the extension period also have not been reported. Funding for the work came from JDRF/Breakthrough T1D and Novo Nordisk ^[2].

Why it matters for patients

For adults with type 1 diabetes who use semaglutide off-label, the results suggest that at least part of the glucose control gained during treatment may be lost within about three months of stopping. That is a practical concern when a prescription ends because of cost, insurance changes, side effects or supply problems.

One reassuring detail: the study found no significant differences between the groups in CGM measures of hypoglycemia, so the worsening numbers were not driven by more low blood sugar ^[1][2]. The decline showed up mainly as less time in the target range and more swings in glucose.

The evidence base is small. With 16 and 22 people in the two groups, the size of any individual's change cannot be predicted from this analysis, and the authors limit their conclusions to adults with type 1 diabetes and obesity using automated insulin delivery. The study also did not test any strategy for tapering, stopping or replacing the drug, and no such guidance comes from this paper.

What happens next

The extension paper was published August 31, 2026 ^[1][2]. The main ADJUST-T1D results appeared earlier in NEJM Evidence ^[2]. No longer follow-up of these participants beyond the 12-week extension has been reported.