GLP-1 Extension Studies: What Happens After a Trial Ends

A GLP-1 extension study is a phase that follows some or all participants after the main trial's planned end, either to see what happens when treatment stops, to keep treating them for longer, or to randomly switch some of them to placebo. These studies are how we know, for example, that most people regain much of their weight after stopping, but each design comes with fine print that changes what it can prove.

Think of the main trial as the movie and the extension as the post-credits scene. Sometimes it reveals the most important plot twist. Sometimes it is a teaser that raises more questions than it answers.

What are the different kinds of GLP-1 extension studies?

There are three main designs, and GLP-1 research has a famous example of each.

Design	What happens	Real example	Strongest at answering
Off-treatment extension	Treatment stops for everyone; a group is followed afterwards	STEP 1 extension	"What happens to people after they stop?"
Randomized withdrawal	Everyone takes the drug first, then people are randomly assigned to continue or switch to placebo	STEP 4, SURMOUNT-4	"Is it the drug that keeps the weight off?"
Long-term continuation	Some participants stay on their assigned treatment for longer	SURMOUNT-1 prediabetes group (176 weeks)	"Does the effect hold on treatment over years?"

What did the STEP 1 extension show?

STEP 1 gave 1,961 adults semaglutide 2.4 mg or placebo for 68 weeks. At week 68, treatment and the lifestyle program were stopped. An off-treatment extension then followed a representative subset of participants for another year.

The headline: one year after stopping, the semaglutide group had regained about two-thirds of the weight they had lost. In numbers, they had lost 17.3% by week 68, regained 11.6 percentage points by week 120, and ended with a net loss of 5.6% from where they started. Most cardiometabolic improvements (like blood pressure and blood sugar markers) drifted back toward baseline too.

Now the fine print, which the authors were upfront about:

• Only 327 people were included. They came from sites in certain countries chosen ahead of time, not the whole trial.
• They lost more than average while on treatment. 17.3% versus 14.9% for the full trial. People who complete 68 weeks of treatment are a slightly different group from everyone who started.
• All analyses were exploratory. There was no pre-planned hypothesis test.
• Both the drug and the lifestyle program stopped together. So it cannot separate the effect of losing the medicine from losing the support.

None of that makes the finding wrong. It tells you how confidently to hold it. We cover the practical side in what happens when you stop semaglutide.

Why are randomized withdrawal trials so useful?

Here is the problem with off-treatment follow-up: there is no fair comparison group still taking the drug. A randomized withdrawal trial fixes that.

STEP 4 (semaglutide). Everyone took semaglutide for a 20-week run-in, reaching 2.4 mg, and lost an average of 10.6%. Then 803 people were randomly assigned: two-thirds continued semaglutide and one-third switched to placebo, for 48 more weeks. Those who continued lost a further 7.9%; those switched to placebo regained 6.9%.

SURMOUNT-4 (tirzepatide). 783 people took tirzepatide for a 36-week open-label lead-in, then 670 who completed it (average loss 20.9%) were randomized to continue or switch to placebo for 52 weeks. Continuing: a further -5.5%. Switching: +14.0%. At week 88, 89.5% of those who continued had kept at least 80% of their lead-in weight loss, versus 16.6% on placebo.

Because people were randomized after responding, the difference between groups is clean evidence that ongoing treatment, not just habits picked up during the run-in, is doing much of the work.

The fine print here is different: these trials only randomize people who got through the lead-in. In SURMOUNT-4, 113 people did not make it to randomization. In STEP 4, 89% reached the full dose and were randomized. So the results describe people who tolerated and responded to the drug, which is exactly the population that faces the "should I stop?" question, but not everyone who ever starts.

What does a long-term continuation tell you?

SURMOUNT-1 kept its 1,032 participants with prediabetes on their assigned dose (or placebo) for 176 weeks, then added a 17-week off-treatment period. Weight change at 176 weeks was up to -19.7% on the highest dose, and progression to type 2 diabetes was 1.3% versus 13.3% on placebo. After 17 weeks off, those numbers were 2.4% versus 13.7%. Because it stayed randomized and blinded through the extended period, this design keeps the fair comparison; its limit is that it covers one subgroup. Our trial duration guide puts it alongside the other timelines.

Five questions to ask about any extension result

1. Who was followed? All participants, a subset, or only completers?
2. Was there still a comparison group, and was it randomized?
3. What exactly changed? Drug stopped, drug continued, dose changed, lifestyle program stopped?
4. How was missing data handled? People drop out of long follow-ups (see why trials report dropouts).
5. Was the analysis planned or exploratory?

Get those five answers and you can usually tell whether an "after the trial" headline is a firm finding or a useful hint.

What happens to participants when a GLP-1 trial ends?

This is the human side of extensions. In many weight trials, study treatment simply stops at the final visit, as in STEP 1. Some sponsors offer an open-label extension or a path to continue; many do not, and participants go back to their own doctor and insurance. If you are thinking of joining a trial, ask about this before you sign; it is on our consent-form question list.

If you ever do stop a GLP-1, by choice, cost or a supply problem, keep tracking. Logging weight, appetite and symptoms in GLPily for the months after makes it much easier for you and your clinician to see what is happening and decide what to do, instead of relying on memory.

Common questions

What is an open-label extension?

A phase where participants and researchers know who is taking what, usually because everyone receives the active drug. It is useful for long-term safety, but without a blinded comparison group it is weaker evidence about effectiveness.

Does the STEP 1 extension mean everyone regains weight after stopping?

It reports an average across 327 people: about two-thirds of lost weight regained within a year. Individuals varied, and the study could not separate stopping the drug from stopping the lifestyle program.

Which is stronger evidence: STEP 1 extension or SURMOUNT-4?

For the question "does continuing treatment keep weight off compared with stopping?", randomized withdrawal trials like STEP 4 and SURMOUNT-4 are stronger because they keep a randomized comparison.