Stopping GLP-1.

In GLP-1 community discussions, "stopping" has become a charged binary: either you're on the medication or you're off it. Three distinct decisions get filed under the same word, and each carries a different evidence base and a different clinical hand-off.

The first is maintenance dosing, where the goal is long-term continuation, often at a reduced dose or extended interval. The second is a planned break, a bounded pause with a known reason: planned pregnancy, upcoming surgery, or the supply and coverage gaps that have become a regular fact of 2025-2026 life. The third is a full off-ramp, intentional indefinite discontinuation after reaching a goal, with the hope of holding the result through habits alone.

All three are legitimate. What makes the conversation tractable is naming which mode you are actually asking about, because the data that matters and the team you need to talk to are different in each. Conflation leads to the wrong question (am I a person who stays on this forever?) instead of the right one (which mode, for which reason, for how long, with which monitoring?).

Three conversations, not one decision.

The table below is a glance, not a recipe.

Mode 1: maintenance.

Maintenance is the mode with the deepest evidence base for sustained results. The WHO's December 2025 obesity pharmacotherapy guideline frames GLP-1 RAs inside a chronic-care model, and the FDA approves Wegovy and Zepbound for chronic weight management, not weight loss. Real-world cohorts (Wiley DOM 2024) show roughly 80.8% of long-term users on lower-than-maximum doses. Maintenance is a first-class choice, not the version of stopping that didn't quite happen.

Mode 2: planned break.

Planned breaks are time-bounded by definition. The signature is that a re-start condition is known up front: after delivery and lactation, after the surgical window closes, after the shortage or coverage gap resolves. Patients in this mode are usually not "stopping" in the long-term sense, but the same biology applies during the gap.

Mode 3: full off-ramp.

The scenario the STEP 1 extension was designed to evaluate: a deliberate indefinite stop, with no scheduled return to medication. It carries the most variable outcome, partly because the trial and real-world populations look so different.

What the data shows on full discontinuation.

Two findings in the discontinuation literature are both real, and they sit in tension. Holding both is the only honest way to summarize what we know.

The STEP 1 extension.

Wilding and colleagues (Diabetes, Obesity and Metabolism, 2022) reported the trial-evidence anchor for cold discontinuation. After 68 weeks on once-weekly semaglutide 2.4 mg, participants had lost a mean of 17.3% of body weight (vs 2.0% on placebo). One year after withdrawal, at week 120, the semaglutide group had regained 11.6 percentage points of the lost weight, leaving net loss at -5.6% (vs -0.1% for placebo). Cardiometabolic improvements drifted back toward baseline alongside the weight. The authors concluded that the findings "confirm the chronicity of obesity and suggest ongoing treatment is required to maintain improvements in weight and health." In the trial population, roughly two-thirds of the lost weight came back within a year of stopping.

Forty-four studies pooled together.

Budini and colleagues (medRxiv 2025, accepted at Lancet eClinicalMedicine 2026) ran a nonlinear meta-regression of 44 studies of weight regain after stopping GLP-1 RAs. Their pooled estimate: regain plateaus at approximately 75.6% (95% CI 68.5-82.7) of the weight lost on treatment, with a regain-curve half-life of about 23 weeks. The shape is fast at first, then slowing, settling well above the weight achieved on therapy but somewhat below pre-treatment weight. The shape matters more than the single number: regain decelerates, but it settles much closer to pre-treatment weight than to your low.

Why real-world cohorts look different.

Gasoyan and colleagues (Diabetes, Obesity and Metabolism / Obesity 2025) followed 7,938 adults in Ohio and Florida who started injectable semaglutide or tirzepatide and discontinued within 3 to 12 months. Mean weight loss before stopping in obesity-treatment patients was -8.4%, lower than trial figures because most patients weren't on a maximally titrated dose for the full duration. Mean weight regain one year after discontinuation: +0.5%. Almost flat, on average.

Two numbers that look incompatible aren't, once you read the methods. Many real-world "discontinuers" quietly restart the same medication, switch to another anti-obesity agent, or were never on a maximally titrated dose to begin with. Real-world "stopping" is leakier than trial-protocol stopping. The trial captures what happens when stopping is genuinely clean; the real world captures what happens when stopping is one move in a longer sequence.

In the trial population, two-thirds of the lost weight came back within a year of stopping. In real-world cohorts, the average regain is much smaller, partly because real-world "stopping" is leakier. Many people quietly restart.

Maintenance: a mode, not a fallback.

The companion article on the maintenance dose walks the mechanics of stepping down. Wu and colleagues (Obesity 2025) published a PK-PD model plus two case patients self-administering tirzepatide every 10 to 14 days; the modeling suggests that reducing dose frequency does not commensurately reduce the weight effect. Modeling plus a case series, not RCT evidence, and the authors are explicit about that. A separate 2,700-patient real-world cohort reported a mean weekly semaglutide dose of 1.08 mg with mean weight loss of 16.7%. Most long-term users are not at label maximum.

Maintenance is the mode with the deepest evidence base for sustained results, not the consolation prize for people who "couldn't stop."

The clinical individualization is the point. There is no single step-down schedule that fits everyone, and we are not going to publish one. The right floor depends on starting dose, length of treatment, weight history, and patient preference; the conversation belongs with the prescriber.

Planned breaks: pregnancy, surgery, supply gaps.

This is the section the companion articles don't cover, because the planned break is structurally different from a maintenance question or a discontinuation question. The reason for the gap defines the rules.

Pregnancy and conception planning.

The FDA labels for semaglutide (Ozempic, Wegovy) recommend discontinuation at least 2 months before a planned pregnancy, owing to the ~1-week half-life. The tirzepatide label (Mounjaro, Zepbound) recommends discontinuation at least 1 month before planned pregnancy. The labels exist because animal reproductive studies showed embryofetal mortality and structural abnormalities at clinically relevant exposures.

A 2025-2026 systematic review of roughly 49,000 GLP-1-exposed pregnancies, paired with the 2022-2025 FAERS disproportionality analysis, did not detect a statistically significant association with major adverse fetal, obstetric, or labor outcomes (one small signal for renal malformations, framed as hypothesis-generating). The authors call this cautious reassurance after inadvertent exposure, not an endorsement of use during pregnancy. Two threads to hold separately: standard practice is still pre-conception discontinuation at the labeled interval, and that guidance has not changed. The post-hoc reassurance is for the person who got pregnant without planning to. Planning belongs with both the prescriber and the obstetric team, ideally before conception.

Surgery prep: guidance that changed in 2024.

The peri-operative story is the cleanest example of why "the last article I read" can be out of date. The 2023 American Society of Anesthesiologists (ASA) consensus statement recommended withholding daily GLP-1 RAs the day of surgery and weekly GLP-1 RAs for one week before an elective procedure, because of delayed gastric emptying and aspiration risk under anesthesia. On October 29, 2024, ASA published a multi-society updated statement with the American Gastroenterological Association, the American Society for Metabolic and Bariatric Surgery, the International Society of Perioperative Care of Patients with Obesity, and SAGES. The update softens the previous blanket recommendation: most patients can continue their GLP-1 RA before surgery, with shared decision-making between the patient, prescriber, anesthesiologist, and surgical team. Patients at highest GI risk are advised to follow a 24-hour liquid diet, or take other case-specific measures.

The international picture (SPAQI consensus, UK and European anaesthesia statements) is converging on the same theme: individualize, don't blanket withhold. The route forward is a conversation with your surgical and anesthesia teams, not a search bar.

Supply and coverage interruptions.

The most common "stopping" trigger in 2025-2026 is the one people don't choose. Cost, formulary changes, manufacturer shortages, and the shifting regulatory status of compounded formulations all produce unplanned pauses. No single formal guidance document exists, but the biology and cohort data both apply: the ~1-week half-life means therapeutic plasma levels persist for several weeks, and meaningful regain sits on a timeline of months rather than weeks. A two-to-four-week gap is not equivalent to discontinuation. The practical move is to re-engage the prescriber early about alternatives rather than waiting it out indefinitely. If a restart is on the horizon, the patterns described in starting again still apply: re-titration usually mirrors the initial titration.

The off-ramp realistically.

For readers choosing or inside Mode 3, the literature describes patterns rather than recipes. Real-world studies describe slow stepwise reduction with monitoring as associated with better weight maintenance at six months than cold-stop; the Wu 2025 paper offers frequency reduction as an alternative or adjunct. Exact interval lengths and total taper duration vary enormously between patients, and that variation reflects how different the starting points are.

What you can expect to track during the transition, regardless of the pattern your clinical team chooses:

• Weight stability, on a smoothed trend, not a daily reading.
• Appetite and food-noise return, ideally on the same scales used during active treatment.
• Lean-mass support, with adequate protein and resistance training during the period when weight is most likely to shift.
• Cardiometabolic markers that improved on therapy. Blood pressure, glucose, and lipids tend to drift back toward pre-treatment ranges in parallel with weight; routine lab follow-up matters.
• Mood and sleep. The FDA in January 2026 asked for removal of the suicidal-ideation warning from GLP-1 labels after a meta-analysis of ~108,000 participants showed no increased risk; mood monitoring during any significant medication change is still appropriate.

There is no single tapering schedule that fits everyone, and we are not going to publish one. The clinical individualization is the point.

Tracking is information, not a verdict. The side-effects timeline article covers a similar logic for symptoms during titration; the same idea applies in reverse during de-escalation.

The chronic-disease reframe.

The framing is not new and it is not controversial inside medicine. The American Medical Association classified obesity as a disease in June 2013 (House of Delegates, Resolution 420). The 2013 AHA/ACC/TOS guideline was the first major guideline to apply a chronic-disease management model to obesity care. The WHO's December 2025 guideline on GLP-1 therapies for adult obesity issues a conditional recommendation for liraglutide, semaglutide, and tirzepatide as part of a "chronic care model." The FDA approves Wegovy and Zepbound for "chronic weight management," not "weight loss," and the ADA's 2024-2025 Standards of Care embed GLP-1 RAs under the same chronic-disease model.

A quiet comparison to hold in mind:

Obesity has been classified as a disease since 2013. The WHO published its first GLP-1 obesity guideline in December 2025. The FDA labels Wegovy and Zepbound for "chronic weight management," not "weight loss." Stopping a chronic-disease medication is a clinical conversation, not a graduation.

The reframe is not that you have to stay on the medication. The reframe is that the GLP-1 did not fix obesity; it changed the trajectory while it was being taken. The off-ramp remains a legitimate choice, but it is one of three modes, not the default goal the other two are deviations from. Nobody calls a statin or an ACE inhibitor the easy way out, and the language around GLP-1 is catching up to a clinical model the rest of medicine has had for years.

An off-medication mode, briefly.

The Symptra changelog lists an "off-medication mode" as a Next-up item, focused on weight stability, habit retention, and a user-set return-to-medication prompt rather than the active-treatment screens. Noted so readers facing one of the three modes know the product thinking exists. Not shipping yet, not a feature pitch.

Red flags worth a call to your prescriber.

During a break or an off-ramp, a handful of patterns are worth a non-urgent message to the prescribing team. None are diagnostic criteria; they are reasons to start the conversation.

• Rapid regain. More than about 1 kg per week sustained over several weeks, beyond normal fluctuation.
• Severe rebound food noise. A return to pre-treatment levels within weeks of stopping, especially if it disrupts work or sleep.
• Mood deterioration. New or worsening depression, anxiety, hopelessness, or suicidal ideation.
• Trouble maintaining adequate protein during the transition, with risk to lean mass during regain.
• Unintended pregnancy while still on a GLP-1 RA, or inside the labeled pre-conception washout.
• Cardiometabolic markers returning to pre-treatment ranges (blood pressure, glucose). The STEP 1 extension showed this happens in parallel with regain; routine lab follow-up matters.

None of the three modes is right for everyone. Stopping is not a binary, and it is not a graduation. It is a clinical decision with three legitimate shapes.

Stopping, reducing, or pausing a chronic medication is a clinical decision that belongs in a conversation with your prescribing team. This article describes what the literature shows; it does not describe what you should do. Specific dose changes, the timing of a pre-conception washout, peri-operative hold decisions, and any restart after a supply gap all belong with the clinicians who know your history. For how Symptra reads and weighs sources, see our editorial process.