Weight Regain After Stopping GLP-1 Drugs: What Trials Found
Published · Page updated
Bookmark this page
Save this article in your browser so you can find it again:
- Computer: Press Ctrl+D (Windows) or ⌘+D (Mac).
- iPhone or iPad: In Safari, tap Share, then Add Bookmark.
- Android: In Chrome, open the menu and tap the star.
Other browsers may use a different menu.
Weight regain was common after treatment withdrawal in major semaglutide and tirzepatide studies. The size of the average regain depended on the study design, follow-up and reference weight. These results support studying maintenance carefully; they do not mean every person regains everything or establish a personal stopping schedule. STEP 4; SURMOUNT-4.
Publisher disclosure: Bio Research LLC operates Bio Peptides Labs and sells research materials. This commercially interested, AI-assisted article is not an independent scientific review. Selected original sources were checked November 14, 2026. Access limits are stated below. Findings for controlled trial formulations do not establish that research catalog materials are suitable substitutes for prescription medicines.

Three studies, three different comparisons
The table summarizes trial findings rather than forecasts for an individual. A plus sign indicates weight gain during the stated interval; a minus sign indicates loss.
Study | Design and comparison | Reported change | Reference point |
|---|---|---|---|
STEP 1 extension | Exploratory follow-up after semaglutide and structured lifestyle support stopped | Semaglutide group regained 11.6 percentage points over one year | Original trial baseline weight |
STEP 4 | After a run-in, continue semaglutide or switch to placebo | −7.9% versus +6.9% over 48 weeks | Weight at randomization, week 20 |
SURMOUNT-4 | After a run-in, continue tirzepatide or switch to placebo | −5.5% versus +14.0% over 52 weeks | Weight at randomization, week 36 |
Sources: STEP 1 extension, STEP 4 full text and SURMOUNT-4 full text.
What “two-thirds regained” actually described
The STEP 1 extension included 327 participants from selected sites. Those previously assigned semaglutide had lost an average 17.3% by week 68. One year after withdrawal, their average net loss from the original baseline was 5.6%. The authors described the regain as roughly two-thirds of the prior loss. Importantly, the trial's lifestyle intervention ended too, and extension analyses were exploratory. Original report.
That is not a statement that everyone regained two-thirds, or that participants ended heavier than they started. It also cannot isolate medication withdrawal from the simultaneous removal of structured support. The headline is more useful when those design limits remain attached to it.
What randomized withdrawal adds
STEP 4 randomized 803 people after a 20-week semaglutide run-in. Both randomized groups continued lifestyle intervention. Its primary comparison therefore asked what happened after continuing treatment versus replacing it with placebo, following the same initial treatment period. STEP 4 methods and results.
SURMOUNT-4 enrolled 783 people into an initial tirzepatide period and randomized 670 after 36 weeks. Those who reached randomization had already experienced substantial weight loss and tolerated the initial treatment. The trial excluded diabetes. Its results describe that selected population, rather than every person who might start tirzepatide. SURMOUNT-4 participant flow and limitations.
Randomization strengthens the comparison between the assigned groups. Selection before randomization still matters when deciding how widely the results apply. Those two points can both be true.
Why percentages cannot simply be subtracted
A percentage needs a denominator. Consider an original arithmetic example: a hypothetical starting weight of 100 units falls by 20% to 80 units. A subsequent 10% gain from 80 is 8 units, producing 88—not 90. The reference weight changed between calculations.
This example is not patient data or an expected treatment response. It explains why a trial's overall baseline-to-end change should not be reconstructed by casually subtracting a loss percentage from a regain percentage. Use the paper's stated time points and methods, and distinguish percentage points of original weight from percent of the weight lost.
Our study-reading guide offers a broader framework for comparing trial estimates without changing their meaning.
New maintenance research asks a more detailed question
The 2026 SURMOUNT-MAINTAIN trial randomized 378 participants after a 60-week tirzepatide period. It compared continued treatment, a protocol-defined reduction and placebo over another 52 weeks. Rescue treatment was available for qualifying regain, and the primary analysis used explicit assumptions about that rescue. It reported better preservation of weight reduction in the active-treatment groups than placebo. Original study abstract.
This is a controlled maintenance experiment, not a general tapering guide. A result that incorporates rescue-treatment assumptions should not be described as a simple uninterrupted year off medication. We rechecked the indexed abstract and its correction record on November 20, 2026; the correction text was still not accessible for review. We therefore do not reproduce the study’s numerical treatment estimates here. Correction record.
Maintenance research should measure more than weight
Future studies should examine function, symptoms, adverse effects, treatment burden and the durability of metabolic changes alongside body weight. They should describe who stops, why they stop and how outcomes are collected afterward. Otherwise, an analysis limited to people who remain in follow-up may miss important experiences.
These are editorial research priorities. They are not claims that Bio Peptides Labs conducts such trials. Our muscle-preservation article discusses the separate distinction between tissue measurements and physical function.
Frequently asked questions
Does everyone regain all the weight after stopping?
No universal outcome is established. The reports show group averages and distributions within specific designs. They cannot predict exactly what will happen to one person.
Do these studies establish a safe taper or restart schedule?
No. This article does not provide one. Decisions about prescribed treatment require a qualified clinician's assessment, including the reason for treatment and any adverse effects.
Is a weight-loss plateau the same as regain after stopping?
No. Continued-treatment plateaus and withdrawal outcomes are different research questions. See GLP-1 weight-loss plateaus for the distinction.
Review method and source access
Checked November 14, 2026. We reviewed relevant full-text methods and results for STEP 4 and SURMOUNT-4, and the STEP 1 extension through its author-institution repository. SURMOUNT-MAINTAIN access was limited to its indexed abstract and correction metadata. This selected-source educational review is not a systematic review or a treatment protocol.
References
On November 20, 2026, we rechecked the four indexed abstracts and the correction's indexed record. The correction text was not accessible; numerical SURMOUNT-MAINTAIN treatment estimates remain omitted. Prior full-text checks of STEP 4 and SURMOUNT-4 retain their November 14 scope; participant-level datasets were not independently reviewed.
