Skip to content
GLP Loss
← Research
Evidence

Tapering Off a GLP-1: What Has Actually Been Tested

One randomized trial has compared lowering the dose against stopping, and its step-down arm never stepped down to zero. A quarter of those reduced to 5 mg still needed rescue therapy, against 8% who stayed at full dose and 67% on placebo.

Owen Castellanos9 min read
The only randomized step-down378 adults randomized at week 60, followed to week 112Weight change from baseline at week 112Continued at max dose-21.9%Reduced to 5 mg-16.6%Switched to placebo-9.9%Received rescue therapy after regaining halfContinued at max dose8%Reduced to 5 mg25%Switched to placebo67%Every arm either continued treatment or stopped outright.No published trial has tested a taper that reaches zero.

Most people who ask how to taper off a GLP-1 are asking two questions at once. Does stepping the dose down soften the landing compared with stopping, and does the landing eventually end somewhere other than the starting weight? One of those now has a randomized answer. The other has none at all, and the gap between them is the most useful thing on this page. What a clean stop produces is measured in the withdrawal article.

What the word taper is doing in this class

In most of medicine a taper is a schedule with an end: a decreasing sequence that terminates in no drug. Applied to obesity pharmacotherapy the word has quietly come to mean something else — a step down to a lower maintenance dose that continues indefinitely, or a stretch of the interval between injections. Those are different interventions with different evidence behind them, and only one of them has been randomized.

The trial that lowered the dose instead of stopping

SURMOUNT-MAINTAIN is a 112-week phase 3b trial run at 20 United States sites. Adults with obesity, or overweight with a weight-related comorbidity, took open-label tirzepatide at their maximum tolerated dose — 10 mg or 15 mg — for 60 weeks. 441 entered that phase and 378 were then randomized 3:3:2 to continue at the maximum tolerated dose, to reduce to 5 mg, or to switch to placebo for a further 52 weeks.[1]

At week 112, the modeled change in body weight from baseline was −21.9% (95% CI, −23.5 to −20.3) on the maximum tolerated dose, −16.6% (95% CI, −18.0 to −15.1) on 5 mg, and −9.9% (95% CI, −11.1 to −8.8) on placebo. Both active arms beat placebo at P < 0.0001, with estimated treatment differences of −12.0 percentage points for the full dose and −6.6 for the reduced one.[1] Ninety-one percent of randomized participants completed.

Read only that and stepping down looks close to free. The reduced arm kept two-thirds of the weight the full-dose arm kept, on a third of the drug.

The rescue column is the finding

From week 84 the protocol allowed rescue tirzepatide for anyone whose regain exceeded 50% of what they had lost. Rescue was given to 11 of 138 (8%) continuing at the maximum tolerated dose, 35 of 142 (25%) on the reduced 5 mg dose, and 60 of 90 (67%) on placebo.[1]

That is a tripling of the rescue rate produced by a dose reduction that cost only 5.3 percentage points of mean weight. The two columns disagree about how benign the step down was because they are measuring different things: the mean describes the group, and the rescue rate counts the individuals for whom the reduction did not hold. The trial’s own conclusion carries the caveat — reducing to 5 mg might provide a valuable alternative to discontinuation, although individuals’ treatment response might vary.[1] A quarter is a lot of variation. The trial was funded by the manufacturer.

The other thing the trial does not say about itself is that its de-escalation arm never de-escalated. It stepped down once and stayed there for a year. Nobody in SURMOUNT-MAINTAIN was tapered off anything.

What stopping outright produces, for scale

The withdrawal evidence is older and consistent. Randomized withdrawal after 20 weeks of semaglutide and after 36 weeks of tirzepatide both produced substantial regain against continued treatment.[2][3] The longest off-treatment follow-up comes from the STEP 1 extension, which tracked 327 participants for a year after both drug and lifestyle intervention were withdrawn at week 68. Mean loss on semaglutide had been 17.3%; participants regained 11.6 percentage points of it, finishing at a net 5.6% below where they started, with most cardiometabolic improvements reverting toward baseline.[4]

Two-thirds back in twelve months, and a third retained. Neither figure is zero, and neither is durable success.

The regain is front-loaded, which is what makes a taper plausible

A 2026 review of the discontinuation literature describes the trajectory rather than only the endpoint: regain after stopping is substantial and a large proportion of it occurs early, in the first months after cessation. The mechanistic account it assembles is a mismatch — appetite returning as pharmacological GLP-1 signaling is withdrawn, compensatory rises in orexigenic hormones including ghrelin, and altered incretin balance — arriving all at once when the drug is stopped abruptly.[5]

That is a coherent argument for smoothing the transition, and the review is careful about how far it goes. Randomized evidence suggests reduced-intensity maintenance attenuates regain compared with abrupt discontinuation. Whether a structured tapering strategy can successfully facilitate discontinuation remains unknown and requires prospective evaluation.[5] The plausible mechanism and the missing trial are both in the same paragraph.

The real-world magnitude, and what else reverts

A 2026 narrative review synthesized the randomized withdrawal trials alongside observational cohorts covering more than 289,000 patients. It reports weight regain of 60% to 90% within one year of discontinuation, with parallel reversal of cardiometabolic benefit: modeling suggests glycemic, blood pressure and lipid parameters return toward baseline within approximately 12 months and HbA1c within 12 to 18 months. It also reports that discontinuation inside the first year is associated with increased risks of coronary artery disease and heart failure against continued therapy.[6]

Its framing is the part worth carrying: discontinuation should be treated as a high-risk clinical transition rather than a treatment endpoint, and it states that validated tapering strategies do not exist.[6] That is a review of associations rather than a trial, and the people who stop early differ systematically from the people who do not.

Stretching the interval: thirty people

The other de-escalation people actually attempt is reducing frequency rather than dose. The published evidence for it is one retrospective case series of 30 adults who had plateaued on weekly semaglutide or tirzepatide and moved to reduced-frequency dosing, usually every other week, at their existing dose. Over an average of 36.3 weeks, mean weight went from 87.9 kg before treatment to 74.1 kg at plateau and to 72.4 kg on the reduced schedule (P < 0.01), with total and truncal fat falling and skeletal muscle mass stabilizing.[7]

Thirty patients, no control group, no randomization, and selected by having already done well. The authors present it as support for structured de-escalation as a promising strategy, which is the right weight to give it. What it is not is a comparison against staying on the weekly schedule, and it is a long way from the sub-therapeutic protocols discussed in the microdosing article.

The one labeled sentence that resembles a taper points backward

None of the four approved labels contains a discontinuation schedule. The closest thing is in the Wegovy injection label and it governs restarting: if 2 or more consecutive doses are missed, dosage escalation is to be reinitiated at a lower dosage, to reduce the risk of gastrointestinal adverse reactions.[8]

So the only labeled instruction about lowering a dose exists because coming back up is hard on the stomach, not because coming down is clinically staged. Whoever restarts after a gap pays the escalation cost a second time, which is also a pricing event on any plan that charges by rung — see the titration article.

The census of what nobody has measured

Four questions, all of them the ones people actually ask, and all of them unanswered by any randomized trial.

Whether a taper ending in no drug beats simply stopping. Whether any particular step schedule — how far down, how long at each step — outperforms any other. Whether a taper combined with structured behavioral or exercise support changes the trajectory, which the reviews propose and none has tested against a control. And whether any of it transfers to a compounded preparation, which is not FDA-approved and is not reviewed by the FDA for safety, efficacy or quality before it is dispensed, and which has no pharmacokinetic data of its own at any dose.

An absence stated plainly is worth more than a plan invented to fill it. Two independent 2026 reviews reached the same conclusion in their own words: validated tapering strategies do not exist, and whether structured tapering can facilitate discontinuation is unknown.[5][6]

What this changes about a purchase decision

Chiefly it removes an exit that people budget around. A course of treatment planned as six months of drug followed by a taper is planned against evidence that does not exist, and the one trial that came closest kept every participant on something for two years. Anyone costing this out should cost the maintenance phase, not the loss phase — which is why a price that holds as the dose moves matters more than an opening rate, and why sellers that hold one price are worth identifying early. What is known about treatment over years rather than months is in the long-term article.

The decision to come off a GLP-1 belongs to the person taking it and the clinician who prescribed it, and there is a legitimate version of it: cost, side effects, pregnancy planning, or a judgment that the trade is no longer worth it. What the evidence does not supply is a schedule that makes it land softly. Anyone being sold one is being sold something nobody has tested.

Frequently asked

Is tapering off a GLP-1 better than stopping suddenly?
No trial has compared a taper that ends against an abrupt stop. The closest randomized evidence, SURMOUNT-MAINTAIN, compared continuing at the maximum tolerated dose, reducing to 5 mg, and switching to placebo for 52 weeks — and the reduced arm stayed on 5 mg rather than coming off. Two 2026 reviews state that validated tapering strategies do not exist.
How much weight comes back after stopping?
In the STEP 1 extension, 327 participants who had lost a mean of 17.3% regained 11.6 percentage points of it over the year after withdrawal, finishing 5.6% below their starting weight. A 2026 review synthesizing trials and cohorts of more than 289,000 patients reports regain of 60% to 90% within one year, with blood pressure, lipid and glycemic improvements reverting over roughly 12 months.
Does taking a lower dose preserve the weight loss?
Partly, and less reliably than the averages suggest. In SURMOUNT-MAINTAIN, reducing to 5 mg produced a 16.6% loss from baseline at week 112 against 21.9% for continuing at full dose and 9.9% for placebo. But 25% of the reduced-dose group needed rescue therapy for regaining at least half their loss, against 8% of those who continued.
What about injecting every other week instead?
The published evidence is a single retrospective case series of 30 adults who had already plateaued on weekly dosing and moved to reduced-frequency dosing at the same dose. Over an average of 36.3 weeks their mean weight fell from 74.1 kg to 72.4 kg with muscle mass stable. There was no control group and no randomization.
Do the labels say how to come off these drugs?
None of them does. The only labeled instruction that lowers a dose is in the Wegovy label and concerns restarting: after two or more consecutive missed doses, escalation is to be reinitiated at a lower dosage to reduce gastrointestinal adverse reactions. Coming off is a clinical decision with no labeled schedule behind it.

Sources

  1. [1] Horn DB, Aronne LJ, Wharton S, et al. (2026). Tirzepatide for maintenance of bodyweight reduction in people with obesity in the USA (SURMOUNT-MAINTAIN): a multicentre, double-blind, randomised, placebo-controlled trial. Lancet. PMID 42119587
  2. [2] Rubino D, Abrahamsson N, Davies M, et al. (2021). Effect of Continued Weekly Subcutaneous Semaglutide vs Placebo on Weight Loss Maintenance in Adults With Overweight or Obesity: The STEP 4 Randomized Clinical Trial. JAMA. PMID 33755728
  3. [3] Aronne LJ, Sattar N, Horn DB, et al. (2024). Continued Treatment With Tirzepatide for Maintenance of Weight Reduction in Adults With Obesity: The SURMOUNT-4 Randomized Clinical Trial. JAMA. PMID 38078870
  4. [4] Wilding JPH, Batterham RL, Davies M, et al. (2022). Weight regain and cardiometabolic effects after withdrawal of semaglutide: The STEP 1 trial extension. Diabetes Obes Metab. PMID 35441470
  5. [5] Damhof MA, van Bruggen FH, van Roon EN (2026). Early Weight Regain After GLP-1 Receptor Agonist Discontinuation: Mechanisms and Implications for Treatment De-Escalation Strategies. Diabetes Obes Metab. PMID 42396734
  6. [6] Shah E, AlShiab R, Abdo A, et al. (2026). Clinical Management of Weight Regain and Cardiometabolic Consequences After Discontinuation of GLP-1 Receptor Agonists. Diabetes Obes Metab. PMID 41889156
  7. [7] Wong M, Wu A, Garhe PK, et al. (2026). Reduced-Frequency GLP1 Therapy Maintains Weight, Body Composition, and Metabolic Syndrome Improvements: A Case Series. Obesity (Silver Spring). PMID 41732031
  8. [8] Novo Nordisk Pharmaceutical Industries, LP (2026). WEGOVY (semaglutide) injection and tablets — Recommendations Regarding Missed Doses DailyMed, U.S. National Library of Medicine. Source

Where to get it

Best GLP-1 injections

Every injectable seller we can verify, with the price each one publishes and an honest read of what the trials measured.

Compare providers →

More in Evidence