Weight Regain After Stopping GLP-1 Medicines: What Trials Show

Weight regain after stopping semaglutide or tirzepatide was common at the group level in the three studies below. Their results describe different follow-up periods and starting weights. They cannot predict how much an individual will regain, or rank medicines against each other.

Sources checked: 26 October 2026. This reference explains the STEP 1 extension, STEP 4 and SURMOUNT-4. It is a focused comparison of their withdrawal designs, not a systematic review of every maintenance strategy.

Three studies, three different starting points

A withdrawal study asks what happens after an initial treatment period. Read the point where treatment changes before reading the final percentage. Tirzepatide acts on both GIP and GLP-1 receptors; semaglutide is a GLP-1 receptor agonist. The GLP-1 foundations guide explains these terms.

What did the withdrawal periods show?

STEP 1: what does “two-thirds regained” mean?

In the 2022 extension paper, the semaglutide arm had lost an average 17.3% by week 68. At week 120, reported net weight loss was 5.6% from the original baseline. The authors described the regain as approximately two-thirds of the earlier loss.

This was exploratory follow-up of a subset of the original trial, with site and participant selection limitations. Both medicine and structured lifestyle intervention ended at week 68. It therefore cannot isolate withdrawal of the medicine from withdrawal of that support.

What do STEP 4 and SURMOUNT-4 add?

STEP 4, published in 2021, randomly assigned people who had completed semaglutide run-in to continue it or switch to placebo. Lifestyle intervention continued in both groups. The resulting comparison addresses continuation versus withdrawal among eligible run-in completers.

SURMOUNT-4, published online in 2023, used an analogous withdrawal question after tirzepatide lead-in, with lifestyle counselling throughout. Its randomised participants were people who completed that initial period. Both trials studied adults with obesity or overweight and a weight-related condition, without diabetes. These selected populations do not represent every person starting treatment.

Why are percentage points and percentages different here?

The denominator is the weight used as the reference. A percentage-point change in loss from original baseline differs from a percentage increase measured after weight has already fallen.

Illustrative arithmetic, not trial data: start at an index of 100, fall to 80, then rise to 90. That is a regain of 10 percentage points of original weight, a 12.5% rise from the lower weight, and half of the earlier 20-unit loss. All describe the same movement.

That distinction explains why “11.6 percentage points regained” must not become “11.6% of lost weight regained”. Always retain the study’s reference time and unit. The trial-reading guide covers endpoints and analysis choices.

Can these results identify the better medicine for maintenance?

No direct drug ranking follows from this table. Each result answers a within-study question. Differences in the initial period, eligibility, follow-up, support and analysis prevent a simple comparison of the regain percentages.

The SURMOUNT-5 head-to-head guide addresses a separate comparison during treatment. It does not convert these withdrawal studies into a head-to-head maintenance trial.

What can a reader reasonably conclude?

• Keep the study question attached to the number. Stopping, continuing, switching and starting treatment are different comparisons.
• A group mean is not a personal forecast. It does not tell every reader the amount or timing of future weight change.
• The observation window matters. Follow-up ending at a given week does not establish what happens indefinitely.
• Weight is one endpoint. These figures alone cannot determine an individual’s overall balance of benefits and harms.

This reference explains evidence rather than a stopping, tapering or restarting protocol. Individual medication decisions require the prescribing clinician’s assessment.

Primary sources and update scope

1. Wilding and colleagues. STEP 1 extension, Diabetes, Obesity and Metabolism (2022). DOI: 10.1111/dom.14725. Funded by Novo Nordisk.
2. Rubino and colleagues. STEP 4, JAMA (2021). DOI: 10.1001/jama.2021.3224. Funded by Novo Nordisk.
3. Aronne and colleagues. SURMOUNT-4, JAMA (online 2023; issue 2024). DOI: 10.1001/jama.2023.24945. Funded by Eli Lilly.

Future updates should identify the new primary study and distinguish a new trial from a later analysis of the same participants. This source check does not imply a named medical review or an exhaustive search.