What Happens When You Stop Zepbound: The Off-Ramp Evidence for Tirzepatide

SURMOUNT-4 is the trial that asked the question directly: what happens if you take tirzepatide for 36 weeks and stop. The answer, in one number: an average 14.0% body-weight regain over the next 52 weeks. Patients who continued lost a further 5.5% in the same window. That dataset is the spine of every reasonable off-ramp conversation about Zepbound. The sibling article on semaglutide is what happens when you stop Wegovy.

What SURMOUNT-4 Showed

SURMOUNT-4 enrolled 783 adults with obesity (or overweight plus a weight-related comorbidity, excluding type 2 diabetes). All received a 36-week open-label lead-in: titration to a maintenance dose of 10mg or 15mg of tirzepatide weekly. The 670 participants who completed the lead-in were then randomised 1:1 to continue tirzepatide or switch to placebo for 52 more weeks.

Over that 52-week randomised phase, the placebo group regained 14.0% of body weight; the continued-tirzepatide group lost a further 5.5%. Across the full 88-week study, mean total weight reduction was 25.3% on continued treatment and 9.9% on placebo (Aronne et al., JAMA 2024; doi:10.1001/jama.2023.24945).

Two details worth noting. The lead-in itself produced a mean 20.9% weight loss, comparable to SURMOUNT-1 at a longer timepoint, so the regain is measured from a deep nadir, not a partial response. The trial also does not measure anything past week 88, roughly a year past randomisation, which is short relative to the chronic-disease framing under which the drug is prescribed.

Why the Tirzepatide Rebound May Differ From Semaglutide Rebound

The data does not answer the comparison cleanly. SURMOUNT-4 reported 14.0% regain at 52 weeks post-randomisation; STEP-4 reported substantial regain after stopping semaglutide at 68 weeks, but with a different lead-in length and baseline. Cross-trial percentage comparisons mislead when the designs differ that much.

What we can say from the pharmacology: tirzepatide produces larger absolute weight loss. SURMOUNT-1 (Jastreboff et al., NEJM 2022) reported a mean 22.5% loss at 15mg over 72 weeks, versus 14.9% for semaglutide 2.4mg in STEP-1. A patient who lost 60 pounds on tirzepatide is putting back 14% of a larger absolute number. In pounds-back-on terms the tirzepatide regain can be larger even at a similar percentage figure.

The mechanistic question, whether GIP plus GLP-1 dual agonism produces a more or less durable post-treatment effect than GLP-1 alone, is not settled. GIP receptor activation may produce adipose-tissue changes that persist longer; or the larger initial loss may drive stronger metabolic adaptation and stronger rebound. There is no head-to-head withdrawal trial.

One supporting signal: the real-world follow-up of SURMOUNT-CN tracked 152 Chinese adults for 26 weeks after a 52-week course. Week-78 weight reduction was 9.1% (10mg) and 12.3% (15mg) versus 1.8% on placebo. More than half of on-treatment loss was maintained at six months post-cessation, consistent with the broader SURMOUNT-4 pattern.

Glycemic and Cardiometabolic Effects After Discontinuation

A post-hoc analysis of SURMOUNT-4 (Horn et al., JAMA Internal Medicine 2025; doi:10.1001/jamainternmed.2025.6112) examined whether cardiometabolic improvements were preserved after withdrawal.

The pattern is what you would expect if the improvements are weight-driven. Larger weight regain was associated with larger reversal in waist circumference, systolic blood pressure, non-HDL cholesterol, A1C, fasting insulin, and HOMA-IR. The reversal was graded: patients who regained less than 25% of what they had lost kept most of their cardiometabolic gains; patients in the highest regain stratum lost most of them.

The clinical reading: tirzepatide's downstream metabolic benefits are not a permanent reset. They track body weight. If weight returns, blood pressure and A1C return with it, in proportion.

One caveat. SURMOUNT-4 excluded patients with type 2 diabetes. In T2D the glycemic effect has a partly weight-independent component (the incretin amplification of insulin secretion), and the SURMOUNT-4 cardiometabolic readout does not extrapolate directly to a T2D population coming off the drug.

The Clinical Case for Staying On vs Tapering Off

The case for continuing: obesity is framed in current guidelines as a chronic, relapsing condition. SURMOUNT-4, SURMOUNT-CN, and the broader GLP-1 discontinuation literature all show meaningful regain when the drug stops. For a patient who responded well and is tolerating it, the evidence supports continuation as the higher-probability path to keeping the result. Most obesity-medicine specialists treat tirzepatide the way they treat antihypertensives or statins.

The case for stopping or stepping down: not every reason to come off is a failure of resolve. Pregnancy planning is a hard indication; the FDA Zepbound label advises stopping at least one month before a planned conception. Persistent intolerable GI side effects after a competent titration are a legitimate reason to stop or switch. Cost drives a substantial fraction of US patients off regardless of clinical recommendation: LillyDirect cash-pay vials run $299 to $499 per month, and many commercial insurers still do not cover the weight-loss indication. If the trigger is that your compounded tirzepatide was discontinued rather than a clinical choice to stop, our guide to the cheaper continuation paths lays them out first. If cost alone is forcing the decision, our guide to being priced off your GLP-1 covers handling it safely. Completion of a defined goal, followed by a planned lower-intensity approach, is a legitimate choice for patients whose comorbidity profile has improved.

What is not a reason to stop, in our reading of the data: a sense that long-term medication is a moral failing. The SURMOUNT-4 regain curve is not a referendum on the patient's discipline. It is the expected outcome of a chronic-disease drug being withdrawn.

What a Structured Tirzepatide Off-Ramp Looks Like

This section describes what published obesity-medicine practice looks like when an off-ramp is the chosen path. Any actual taper should be done with the prescriber who knows your dose, response, and reasons for stopping. The article is awaiting medical review.

There is no published, randomised, tirzepatide-specific taper schedule. There is also no withdrawal syndrome; the FDA label does not require a taper for safety, and abrupt discontinuation is pharmacologically allowed. The argument for tapering is behavioural: give nutrition habits and non-drug supports time to take over before the satiety signal fully clears.

The pattern that appears most often in obesity-pharmacotherapy guidance is a step-down through the existing FDA-approved dose ladder (15mg, 12.5mg, 10mg, 7.5mg, 5mg, 2.5mg, off), holding each step for four to eight weeks while monitoring weight and appetite. From 15mg, that produces an off-ramp of several months.

Some clinicians use a maintenance-dose strategy instead, holding indefinitely at a lower dose (commonly 5mg or 7.5mg) rather than coming all the way off. That is a different decision; SURMOUNT-4 did not test it.

Non-pharmacological supports consistently associated with reduced post-discontinuation regain in the broader weight-maintenance literature: structured high-protein nutrition (1.2 to 1.6 g/kg/day), resistance training to preserve lean mass, sleep regularisation, and behavioural support of some form.

The Data Gap

Long-term follow-up beyond three years after planned tirzepatide discontinuation is essentially zero in the published literature. SURMOUNT-4's randomised window is 52 weeks past stopping. SURMOUNT-CN's real-world follow-up is 26 weeks. We have no published data on year three or year five.

That is the central tension. The drug is prescribed as chronic-disease treatment; the withdrawal evidence extends roughly one year past the stop. A patient and a prescriber can read the same SURMOUNT-4 readout and reach defensibly different conclusions, because the data runs out before the disagreement is settleable.

What to Discuss With Your Prescriber

1. How does my response curve compare to the SURMOUNT-4 lead-in? A patient who lost 25% in 36 weeks is in a different position from one who lost 8%.
2. Which of my reasons for stopping would change if cost, side effects, or insurance changed? Sorting modifiable from non-modifiable reasons clarifies whether stopping is the goal, or a change in conditions is.
3. If I stop, what is the monitoring plan for weight, blood pressure, and A1C, and at what threshold do we re-evaluate? A taper without a re-evaluation rule is a discharge.
4. Is a maintenance-dose strategy appropriate for me? A separate decision from tapering to zero, with different (thinner) evidence.
5. What is the side-effect picture at lower dose levels? Some patients have more nausea at dose transitions than at maintenance.
6. What is the plan if I regain a meaningful fraction in the first 6 to 12 months? Agree resumption thresholds before the off-ramp begins.

How we keep this article current

We refresh this page on any SURMOUNT extension publication, tirzepatide-specific taper trial, or retatrutide withdrawal data. Two areas tend to drift faster than the rest:

• SURMOUNT extension data. Additional analyses by Aronne and colleagues, plus longer-term follow-up of the original SURMOUNT cohorts, are reaching peer review through 2026. The likely change is at the two-to-three-year horizon, where we currently have almost nothing.
• The retatrutide programme. Lilly's GLP-1 + GIP + glucagon triple agonist had a larger Phase 2 weight loss than tirzepatide, and its Phase 3 read-outs land across 2026 and 2027. If retatrutide is approved and behaves differently on withdrawal, the question of how off-ramp dynamics scale with mechanism class becomes testable.

If you spot an error or missing source, please email [email protected].

For the semaglutide version of the same question, see what happens when you stop Wegovy. For the underlying mechanism, see how GLP-1 drugs work. Our editorial process is on the methodology page; medical-content limits are on the disclaimer.