Research
What Happens When You Stop: The Weight-Regain Evidence
Medically reviewed by Dr. J. Adam, MD · Updated July 23, 2026 · Sources cited below
SURMOUNT-4 ran a 36-week open-label lead-in during which all 670 participants lost weight on tirzepatide, then randomised them to continue or switch to placebo for 52 weeks. Continuing maintained and modestly extended the loss. Stopping produced substantial regain. Because everyone succeeded first, the divergence is attributable to withdrawal itself.
Key takeaways
- Everyone lost weight first, then was randomised — group differences cannot explain the regain.
- Continuing maintained the result; switching to placebo produced substantial regain over 52 weeks.
- Regain reflects returning appetite signalling, not a failure of willpower.
- The finding makes sustained monthly cost the number that matters, not an introductory rate.
- No cited trial establishes that any tapering method prevents regain.
| Trial | SURMOUNT-4 (NCT04660643) |
|---|---|
| Design | Randomised withdrawal after 36-week lead-in |
| Participants | 670 |
| Randomised follow-up | 52 weeks |
| Identifiers | PMID 38078870 · doi:10.1001/jama.2023.24945 |
| Related | SURMOUNT-MAINTAIN (Lancet 2026) tested reduced-dose maintenance |
Why is this design so hard to argue with?
Observational reports of regain after stopping can always be explained away. Perhaps the people who stopped were struggling anyway; perhaps they had less support or more life disruption. Those objections are reasonable and they are why observational data rarely settles anything.
SURMOUNT-4 removes them. Every participant went through the same 36-week open-label lead-in and lost weight on the drug. Only then were they randomly assigned to continue or switch to placebo. At the moment of divergence the two groups were comparable by construction.
That makes the subsequent separation attributable to the withdrawal itself, which is as close to a causal statement as clinical research gets outside of a mechanism experiment.
Why does weight come back?
Because the effect is pharmacological rather than educational. While the drug is present it slows gastric emptying and suppresses appetite signalling through incretin receptor pathways. Remove it and those systems return to baseline over the following weeks as exposure declines.
Appetite returns, intake rises, and weight follows. Nothing about the person's knowledge, habits or motivation has changed — which is exactly what the randomised design demonstrates, since motivation was not the variable being manipulated.
This is why framing regain as relapse is both inaccurate and counterproductive. It discourages honest conversation with prescribers, and people who believe regain is a personal failing are less likely to report that they stopped.
View chart data as a table
| Series | Randomised | Wk 12 | Wk 24 | Wk 40 | Wk 52 |
|---|---|---|---|---|---|
| Continued treatment | 0% | -1.9% | -3.2% | -4.1% | -5.5% |
| Switched to placebo | 0% | 4.6% | 8.3% | 11.2% | 14.0% |
What does this mean for what you spend?
It moves the relevant number from the first month to the sustained monthly rate over years. A programme with a low introductory price and a high renewal price can cost far more across two years than one with a flat, slightly higher rate.
Two years is the conservative planning horizon, not the ambitious one. At the manufacturer's direct cash rate for a maintenance dose, two years of continued treatment runs to roughly $10,800 before any coverage. That is the figure worth planning against.
It also reframes what a good provider looks like. Introductory discounts matter far less than the recurring rate, whether the price holds at higher doses, and what happens to prepaid money if you stop.
View chart data as a table
| Covered benefit with savings card | 600 |
|---|---|
| Self-pay starting dose | 7176 |
| Self-pay maintenance dose | 10776 |
| Retail without coverage | 26064 |
Does everyone regain everything?
No, and the trial does not claim that. Regain was substantial on average and incomplete within the 52-week window, and individual trajectories varied considerably around the group mean.
What the data does not support is the expectation that a limited course produces a durable result. That expectation is common, is reinforced by marketing that presents a finish line, and is the single most consequential misunderstanding in this category.
It is worth separating two different questions. Whether you will regain some weight after stopping is well answered: on average, yes, substantially. How much you specifically would regain, and over what period, is not answerable from group data.
What are the alternatives to stopping?
Until 2026 the honest answer was that nobody had tested one. SURMOUNT-MAINTAIN changed that: reducing to 5 mg maintained 16.6% bodyweight reduction against 21.9% on the maximum tolerated dose and 9.9% on placebo. A reduced dose is now an evidence-supported middle option rather than a hopeful guess.
If cost is what is driving a decision to stop, that is worth raising explicitly, because the options are different: an insurance appeal, a manufacturer savings programme, a different pathway, or a dose reduction now have distinct evidence behind them.
If side effects are driving it, dose adjustment may resolve the problem while preserving benefit. Either way, stopping without that conversation forfeits options that exist, and restarting later is not always simple — supply, coverage and re-escalation all re-enter the picture.
| Option after reaching target | Evidence | Trade-off |
|---|---|---|
| Continue at maximum tolerated dose | SURMOUNT-4, SURMOUNT-MAINTAIN | Highest maintained result; highest cost |
| Reduce to a lower maintenance dose | SURMOUNT-MAINTAIN (Lancet 2026) | Held 16.6% vs 21.9%; usually cheaper |
| Stop entirely | SURMOUNT-4 | Substantial regain on average |
| Taper gradually | No cited randomised evidence | Untested; not published as a schedule here |
Frequently asked questions
Will I regain weight if I stop?
Substantial regain followed randomised withdrawal in SURMOUNT-4. Individual results varied.
How fast does it happen?
Regain accumulated across the 52-week randomised period. The trial does not support a precise personal timeline.
Does tapering prevent it?
No cited trial establishes that. SURMOUNT-4 compared continuing against abrupt substitution and did not test tapering.
Is regain my fault?
No. Participants regained after random assignment to stop, with nothing about their motivation changing at that moment.
Is a lower dose an option?
Yes — SURMOUNT-MAINTAIN found 5 mg maintained 16.6% against 9.9% on placebo.
Sources
Every clinical claim above links to a primary source: an FDA record, a peer-reviewed publication with DOI or PMID, or a ClinicalTrials.gov registration. Where a figure could not be verified against a primary source, it is labelled rather than asserted.
rel="nofollow noopener". Rankings follow our
published methodology. See conflicts
of interest. This article is information, not medical advice.