What Happens When You Stop a GLP-1: The Withdrawal Data, Trial by Trial
Weight returns. In the STEP-1 extension roughly two-thirds of the weight lost was regained within a year of stopping. In SURMOUNT-4, participants withdrawn to placebo regained weight while those who continued kept losing. This is the most consistent finding across the class, and it is the one that turns a price comparison into a maintenance question.
What each withdrawal study reported
| Trial | Design | What happened after withdrawal |
|---|---|---|
| STEP-1 extension | Off-treatment follow-up after 68 weeks | About two-thirds of lost weight regained within a year |
| SURMOUNT-4 | Randomised withdrawal after a lead-in | Regain on placebo; continued loss on treatment |
| STEP-4 | Randomised withdrawal at week 20 | Regain in the withdrawn arm; continued loss on treatment |
Separate trials with different designs and populations. What they share is direction.
Indexed: 149 represents the 14.9% mean loss at the trial endpoint. Read as a direction of travel from the published extension, not as a tabulated value.
Why does it come back?
The honest answer is that the drug is doing something while it is present and stops doing it when it is not. GLP-1 receptor agonists reduce appetite and slow gastric emptying; withdraw the drug and both effects resolve. Obesity is managed on this model as a chronic condition, in the same way blood pressure is: the medicine controls it rather than curing it.
That framing is not a marketing line. It is what the withdrawal data shows, and it is why the label describes chronic weight management rather than a course of treatment.
| Route | Monthly at maintenance | Twelve months |
|---|---|---|
| Covered brand plus savings offer | $25 | $300 |
| Compounded, flat, 12-month plan | $145 | $1,740 |
| Approved oral tablet, maintenance dose | $299 | $3,588 |
| Brand pen, self-pay | $349 | $4,188 |
| Brand pen, higher dose | $399 | $4,788 |
Manufacturer self-pay figures read on the published price guide, 4 August 2026.
What the data does not settle
- Whether tapering helps. No trial has tested a structured taper against abrupt discontinuation for weight maintenance.
- Whether intermittent dosing preserves the benefit. Not studied.
- Who regains fastest. Trial means conceal wide individual variation, and no validated predictor exists.
- Whether behavioural support changes the trajectory. Both arms of these trials received lifestyle intervention, so its separate contribution after withdrawal is not isolated.
Anyone selling a protocol that answers those questions is ahead of the published evidence.
What comes back besides weight
The regain figure is the headline, but the withdrawal data covers more than the scale. Published extensions report that cardiometabolic improvements move back toward baseline alongside weight — blood pressure, lipids and glycaemic measures that improved during treatment.
That matters for how the decision is framed. If someone started treatment for a weight-related comorbidity rather than for weight itself, the question at withdrawal is not only whether the weight returns but whether the reason for treating returns with it.
| Measure | Direction after stopping | Where reported |
|---|---|---|
| Body weight | Returns toward baseline | STEP-1 extension, SURMOUNT-4, STEP-4 |
| Waist circumference | Returns toward baseline | STEP-1 extension |
| Blood pressure | Moves back toward baseline | STEP-1 extension |
| Lipid measures | Move back toward baseline | STEP-1 extension |
| Glycaemic measures | Move back toward baseline | STEP programme |
Directional findings from published extensions. Individual variation is wide.
The commercial structures that punish stopping
Three pricing structures interact badly with a chronic-treatment model, and none of them is disclosed prominently.
- Prepaid multi-month plans. They commit a patient through titration — the exact window where discontinuation for adverse events concentrates. Ask what is refundable before paying.
- Membership fees separate from medication. A membership continues whether or not you are taking the drug, and it is frequently excluded from the advertised monthly figure.
- Dose-tiered pricing. The advertised price is the starter dose. Every published efficacy figure was collected at a maintenance dose, so the advertised price and the price of treatment are different numbers.
The figure that matters when treatment is ongoing. Manufacturer self-pay prices read on the published guide, 4 August 2026.
What a reasonable conversation with a prescriber covers
None of this is a reason to start or continue treatment, and none of it is medical advice. It is a list of things worth raising:
- Whether the original reason for treating still applies.
- What the plan is if cost becomes the constraint — a covered route, a different molecule, a different dose.
- What monitoring makes sense if stopping is the decision.
- That the regain data describes a population mean, and individual trajectories vary widely.
Why "it only works while you take it" is not a criticism
The regain data is sometimes presented as though it exposes the drugs as ineffective. It does not. Antihypertensives do not cure hypertension; statins do not cure dyslipidaemia. In both cases stopping returns the measure toward baseline, and nobody considers that a scandal.
What the data does establish is the model of care: this is chronic management, and it should be priced, planned and discussed as such. The failure mode is not the drug working only while taken — it is a market that advertises an introductory month for a treatment the evidence says is ongoing.
The three numbers worth writing down before starting
| Question | Why it decides more than the headline price |
|---|---|
| What will I pay at a maintenance dose? | Every efficacy figure was collected at a maintenance dose |
| What does twelve months cost? | The withdrawal data says treatment is ongoing |
| What is refundable if I stop in month two? | Discontinuation concentrates during titration |
A programme that will not answer all three in writing has answered the question.
What the withdrawal data does not settle
- Whether a structured taper changes the trajectory. Not tested.
- Whether intermittent or maintenance-only dosing preserves benefit. Not tested.
- Who regains fastest. No validated predictor exists; trial means conceal wide individual variation.
- Whether behavioural support alters the curve after stopping. Both trial arms received lifestyle intervention, so its separate contribution is not isolated.
Anyone selling a discontinuation protocol that answers those questions is ahead of the published evidence, and should be asked which trial they are citing.