GLP-1 Microdosing in 2026: What the Evidence Actually Shows

This is the long version of a question that did not exist three years ago and is now one of the most searched phrases in weight-loss medicine. "GLP-1 microdosing" started as a social-media workaround and became a product category: telehealth programs sell it, clinics advertise it, and a health-tracking company found that one in seven people on an injectable GLP-1 has tried it. The medical societies that study obesity have responded with the same sentence in different words: there is little to no evidence for it.

That sentence is true, and it is not the whole story. The reasons people microdose are real: the side effects at full dose, the price of staying on treatment for years, and the near-universal regain when treatment stops. Each of those problems has an evidence base of its own, and some of it points at lower doses, just not at the fractions of a starting dose that "microdosing" usually means. This guide sets the two apart. It walks through what the word means in practice, who is doing it and why, where the labelled doses come from, what the dose-response trials actually show, the one real-world study of people who stayed at the lowest approved dose, the randomized trial of a reduced maintenance dose, what happens when people stop entirely, what the microdose programs sell and at what price, the compounded-vial problem behind the FDA's dosing-error alerts, and the regulatory position in October 2026.

How to read it. Every clinical figure carries the trial, paper, label or statement it comes from, with its date, linked inline. Figures that reach us through press reporting, conference coverage or company statements rather than a peer-reviewed paper are labelled as reported, with the outlet and date. Every provider price is the one in our verified GLP-1 price index, checked on each provider's own site on October 16, 2026, or a program's published price as captured on the dated page we link to. Nothing is estimated by us, and where the evidence does not exist we say so rather than fill the gap. This guide is general information, not medical advice: whether any dose of any GLP-1 is appropriate for you is a decision for a licensed clinician who knows your history.

What this guide is not

It is not a dosing chart. We do not publish fractional-dose schedules, "units" tables or vial-stretching arithmetic, because no such schedule has been tested and because the FDA's adverse-event reports for compounded GLP-1s are, in large part, reports of people measuring small volumes wrong. If you are on a GLP-1 and want a lower dose, the labelled reductions and the trial-tested maintenance dose are described below, and the right person to set your dose is the prescriber who can see your chart.

The first problem with microdosing is that nobody agrees what it is. The Obesity Medicine Association's Access and Policy Working Group, in a perspective statement published in 2026 (Microdosing of obesity medications, PDF hosted March 2026), calls it "a marketing term used to describe the administration of very small, often subtherapeutic amounts of obesity medications." Jody Dushay, an endocrinologist at Beth Israel Deaconess Medical Center and assistant professor at Harvard Medical School, wrote in STAT on May 29, 2026 that no one has defined what counts as a microdose and that no legitimate long-term data back the trend. Amanda Velazquez, director of Obesity Medicine at Cedars-Sinai, told the hospital's own expert-advice page (published in mid-2026) that it "is a term that has been popularized on social media," not a medical one.

In practice the word covers at least three different things, and it matters which one someone means.

Meaning one: a fraction of the starting dose

The strict sense. The lowest labelled dose of injectable semaglutide for weight management is 0.25 mg a week, and of tirzepatide 2.5 mg a week; both are described in their labels as initiation doses, not treatment doses. A microdose in this sense is below those numbers: a tenth of a milligram of semaglutide, a milligram of tirzepatide, sometimes less, usually drawn from a compounded vial. One consumer guide summarised in a Forbes Health piece on microdosing costs put it as doses "below 0.25 mg weekly, far outside any range studied in clinical trials." This is the version the societies are talking about when they say there is no evidence, because there is none: no trial of either drug has tested a dose below the starting dose for any outcome.

Meaning two: staying on the starting dose

The common sense. Many people who describe themselves as microdosing are taking the labelled starting dose, or the first step up, and simply never escalating. Noom's microdose program, as Reuters reported in August 2025 (carried by PharmaLive), sold compounded semaglutide up to a maximum of 0.6 mg a week, which is a labelled-range dose, not a fraction of one. This version has a sliver of evidence, because the lowest doses were arms in the dose-ranging trials and because a 2026 records study looked at people who stayed there. That evidence is covered below, and it is not encouraging about weight loss.

Meaning three: a labelled dose, less often

The stretched sense. Taking a full dose every ten days or every two weeks instead of weekly is sometimes called microdosing too, because the average weekly exposure falls. A reality-television weight-loss story that circulated in mid-2026 described exactly this, a full injection every few weeks. It is a different pharmacology from a small weekly dose, because both drugs have half-lives of roughly a week and a dose every two weeks produces a sawtooth of high and low blood levels rather than a steady low one. It is also untested.

Why the distinction matters

Each meaning carries a different risk. The first requires measuring very small volumes from a vial, which is the mechanism behind the FDA's dosing-error reports. The second is a labelled dose and the question is only whether it does anything useful. The third produces swings in drug level that the trials never studied. When a program, a clinician or a friend says "microdosing," the first question to ask is which of the three they mean, and the second is whether the dose will be written in milligrams.

The best count comes from Evidation, a health-tracking company with a large app-based research cohort. In a survey fielded in December 2025 and reported by Medscape on January 27, 2026, 14.6% of injectable GLP-1 users said they had taken or were taking doses below the FDA-approved levels. The survey went to 60,157 members, of whom 6,574 were current injectable users, 667 were on the pill, 3,846 were former users and 49,070 had never taken a GLP-1. Evidation described the results as preliminary and said more than one-third of current microdosers were making dosing decisions on their own, without a provider's guidance. About half had started on a standard regimen, and about half were using compounded versions.

A second cut of the same program, presented as a poster at ISPOR 2026 in May 2026, analysed 8,486 current injectable users and found 15.0% reporting microdosing: 9.3% currently, 5.7% formerly. The reasons given, in order, were management of side effects (40.8%), cost reduction (31.9%) and support for weight maintenance (29.6%). Current microdosers were more likely than other users to get their medication from somewhere other than their regular healthcare provider: telehealth, a weight-loss clinic or a medspa. The two Evidation figures come from different samples and should not be merged; the point they share is that roughly one in seven injectable users has tried it, and that the reasons are practical rather than ideological.

The caveat Evidation itself gives is that its cohort is self-selected app users, so the share may not generalise to all US patients. It is the only large count that exists.

The size of the population behind that one in seven

The base is large and growing fast. Gallup's National Health and Well-Being Index, as reported by local news on July 10, 2026, found 11% of US adults currently taking a GLP-1 for weight loss in 2026, up from 3% two years earlier, and 15% saying they had used one at some point, up from 6% in 2024, in a survey of 5,065 adults. EY-Parthenon research reported by CBS News on October 22, 2026 found that about 54% of people on semaglutide drugs now take them mainly to lose weight, more than for diabetes or any other reason. One in seven of 11% of adults is a population in the millions, and the market has noticed.

Three reasons, three different questions

The three motives in the Evidation data map onto three separate evidence questions, and this guide takes them one at a time. Side effects: does a lower dose reduce them, and what does the label already allow a clinician to do about them? Cost: does a lower dose actually cost less, given how the providers we track price their plans? Maintenance: does a lower dose hold weight after it has been lost, and what did the one randomized trial of a reduced maintenance dose find? The honest summary is that the first has a partial yes, the second is mostly a no at flat-price providers, and the third has evidence only for a labelled dose, not a microdose.

To judge a dose below the ladder, you need the ladder. Both approved injectables are started low and raised in steps, for tolerability rather than effect: the step doses were chosen in the development programs as the gentlest way up, and the treatment doses are the ones at the top.

Semaglutide (Wegovy)

The Wegovy label's dosing table, as summarised by Drugs.com from the prescribing information, runs 0.25 mg once weekly for weeks 1 to 4, 0.5 mg for weeks 5 to 8, 1 mg for weeks 9 to 12, 1.7 mg for weeks 13 to 16, then a maintenance dose of 1.7 mg or 2.4 mg from week 17. If a step is not tolerated, the label allows holding the current dose for four more weeks before increasing. If 2.4 mg is not tolerated, the dose can be lowered to 1.7 mg for up to four weeks and then raised again, and treatment is stopped if 2.4 mg is still not tolerated (Ventura County formulary summary of the label). Note what that means: the label already contains a slower ramp and a temporary step down. It does not contain a dose below 0.25 mg, and 0.25 mg is defined as a starting dose.

The ladder gained a rung above in 2026. On March 19, 2026, the FDA approved Wegovy HD, semaglutide 7.2 mg once weekly, for adults with obesity who have tolerated 2.4 mg for at least four weeks and need further weight reduction (Novo Nordisk press release, March 19, 2026). The approval came under the FDA's national priority voucher program, and the company announced nationwide availability on April 7, 2026. The direction of regulatory travel, in other words, is toward higher doses with more effect, not lower ones with less.

Tirzepatide (Zepbound)

The Zepbound label (FDA-hosted prescribing information, 2026 revision) sets the starting dose at 2.5 mg once weekly for four weeks and states in so many words that "the 2.5 mg dosage is for treatment initiation and is not approved as a maintenance dosage." After four weeks the dose rises to 5 mg, and may then be increased in 2.5 mg increments after at least four weeks on the current dose. The recommended maintenance doses are 5 mg, 10 mg or 15 mg once weekly, chosen on response and tolerability. The 2026 label lists six strengths: 2.5, 5, 7.5, 10, 12.5 and 15 mg.

Two things follow. First, "staying on 2.5 mg" is explicitly outside the label, which is why the records study below had to treat it as a proxy for microdosing rather than a prescribed strategy. Second, 5 mg is a labelled maintenance dose, which is why the SURMOUNT-MAINTAIN trial could test it, and why its result is evidence about a reduced labelled dose and not about microdosing.

The pills

The two oral GLP-1s approved in the last year have ladders too: the Wegovy pill steps from 1.5 mg to its 25 mg maintenance dose, and Foundayo across 6, 12 and 36 mg tablets, as set out in our Foundayo vs Wegovy pill guide. Tablets cannot be split into fractional microdoses the way a vial can be drawn, which is one reason the microdosing conversation is almost entirely about injectables, and almost entirely about compounded ones.

The single most useful fact about microdosing is that both drugs have clean dose-response curves, measured in randomized trials. More drug, more weight loss, more side effects; less drug, less of each. Nothing in the trials suggests a floor below which a small dose keeps the benefit and loses the cost.

Semaglutide: the phase 2 dose-ranging trial

Before Wegovy existed, Novo Nordisk ran a 52-week, randomized, double-blind dose-ranging trial of once-daily semaglutide in 957 adults with obesity and no diabetes, published by O'Neil and colleagues in The Lancet in 2018 (volume 392, pages 637 to 649). Five daily doses were tested, 0.05 mg to 0.4 mg, against liraglutide 3 mg and placebo. As reported from the trial and its ENDO 2018 presentation, estimated mean weight loss at one year ran from about 6% on the lowest dose to 13.8% on the highest, against 2.3% on placebo and 7.8% on liraglutide; the share of participants who lost weight rose from 54% in the 0.05 mg group to 83% in the 0.4 mg group. The lowest dose did something, roughly what liraglutide did, and the highest did more than twice as much. The weekly 2.4 mg Wegovy dose was built from the top of that curve, not the bottom.

Tirzepatide: SURMOUNT-1

SURMOUNT-1, the 72-week phase 3 trial behind Zepbound's approval, randomized participants to 5, 10 or 15 mg weekly or placebo. The reported weight changes at week 72 were 15.0%, 19.5% and 20.9% for the three doses against 3.1% for placebo (figures as carried in a 2026 evidence review of microdosing citing the trial; the primary paper is Jastreboff et al., New England Journal of Medicine, 2022). The 5 mg arm is the lowest maintenance dose on the label and the lowest dose in the trial. Nothing lower was studied; 2.5 mg existed only as the four-week on-ramp.

Going the other way: STEP UP and the 7.2 mg dose

The curve keeps rising above 2.4 mg. STEP UP, published in The Lancet Diabetes & Endocrinology in November 2025 (volume 13, issue 11, pages 935 to 948), compared semaglutide 7.2 mg with 2.4 mg and placebo over 72 weeks in adults with obesity. In the trial's primary analysis, which counts everyone randomized, mean weight loss was 18.7% on 7.2 mg, 15.6% on 2.4 mg and 3.9% on placebo, a difference of 3.1 percentage points between the two active doses. Among people who stayed on treatment, the figures were 20.7%, 17.5% and 2.4%, which are the numbers Novo Nordisk quotes. Gastrointestinal adverse events rose with dose too: 70.8% on 7.2 mg, 61.2% on 2.4 mg, 42.8% on placebo. HCPLive's coverage of the approval reported altered skin sensation (dysesthesia) in 22% on 7.2 mg against 6% on 2.4 mg and 0.3% on placebo (HCPLive, March 2026).

What the curve implies for a microdose

Put the three trials in a row and the pattern is monotonic from 0.05 mg daily to 7.2 mg weekly: each step up buys more weight loss and more side effects, and each step down gives back some of both. A microdose sits below the left edge of every measured curve. Extrapolating the curve says it would produce less than the lowest tested dose produced, which for semaglutide was about 6% at one year and for tirzepatide 15%. Extrapolation is not evidence, and the honest statement is that no one has measured it. But the trials give no reason to expect a special property at very low doses that the measured doses lack.

No. As of October 2026 there is no randomized controlled trial of a dose of semaglutide or tirzepatide below the labelled starting dose, for weight loss, weight maintenance, side-effect reduction, cardiovascular outcomes or anything else. Every statement to that effect from a medical society, an academic medical centre or an individual clinician in 2026 is simply accurate. What exists instead is one observational study of people who stayed at the lowest approved dose, which is the closest thing to real-world microdosing data anyone has published.

The 2026 electronic-records study

Murugadoss, Venkatakrishnan and Soundararajan published "Comparison of Cardiometabolic Benefits and Tolerability in Patients on Sustained Low Doses of Semaglutide or Tirzepatide" in Biology Methods and Protocols on August 31, 2026. They used de-identified electronic health records from a federated network of about 29 million patients and defined "sustained low dose" as at least three distinct prescriptions for the lowest approved dose only, spanning at least six months: 0.25 mg for semaglutide, 2.5 mg for tirzepatide. Of 490,072 semaglutide-treated patients, 814 met that definition; of 322,442 on tirzepatide, 1,016 did. The authors state plainly that the records cannot show whether anyone was intentionally microdosing, and that the lowest approved doses are a proxy for the practice rather than the practice itself.

The results, as the abstract reports them: mean weight change at 12 months was 2.2% on sustained low-dose semaglutide and 5.5% on sustained low-dose tirzepatide. Over 24 months, low-dose tirzepatide carried higher rates of several adverse events than low-dose semaglutide, including constipation (32.6% against 22.4%) and acute kidney injury (2.7% against 0.4%). Adverse events were identified with AI-assisted review of clinical notes, and the study is observational, so it cannot establish cause. Healthline's report on the study (September 18, 2026) added a comparison from a co-author: in a matched analysis, patients on higher semaglutide doses lost about 12% against 1.7% on the lowest dose. Healthline also noted that lower doses were linked to fewer side effects, which is the one part of the microdosing case this study supports.

How to read 2.2% and 5.5%

Set them against the full-dose numbers. In the same group's separate PNAS Nexus analysis of full-dose patients in 2026, mean loss was 14.7% on tirzepatide and 10.8% on semaglutide. In the trials it was about 15% on semaglutide 2.4 mg (STEP 1) and up to 20.9% on tirzepatide 15 mg (SURMOUNT-1). The lowest-dose patients lost roughly a fifth to a third as much. Cedars-Sinai's expert page puts it the same way: weight loss at very low doses "tends to be modest, around 5% to 6%," compared with far larger losses at approved doses. That is also roughly what a structured lifestyle program produces without medication, and roughly what liraglutide produced in the 2018 dose-ranging trial.

One more caution. The 814 and 1,016 people in this study were not microdosing in the strict sense. They were taking the labelled starting dose, repeatedly. A fraction of that dose, drawn from a vial, has no published outcome data of any kind, and this study cannot be stretched to cover it.

No, and the confusion between them is where most of the apparent evidence for microdosing comes from. A maintenance dose is a labelled dose, taken at the labelled interval, after the weight has been lost. The first randomized evidence that a reduced maintenance dose can hold weight arrived in May 2026, and it is the strongest argument that "less drug after the loss" is a real strategy. It is not evidence for a microdose, because the dose it tested was 5 mg of tirzepatide, twice the starting dose and the lowest maintenance dose on the label.

SURMOUNT-MAINTAIN

Published in The Lancet in May 2026 and summarised by the American College of Cardiology on May 18, 2026, SURMOUNT-MAINTAIN took people who had lost weight on tirzepatide over a 60-week lead-in and randomized 378 of them, 3:3:2, to continue their maximum tolerated dose (10 or 15 mg), drop to 5 mg, or switch to placebo for a further 52 weeks: 140, 144 and 94 participants respectively. At week 112 the model-based weight change from the original baseline was 21.9% below baseline on the full dose, 16.6% on 5 mg and 9.9% on placebo; the differences against placebo were 12.0 and 6.6 percentage points. Rescue therapy, offered to anyone who regained at least half of what they had lost, was needed by 8% on the full dose, 25% on 5 mg and 67% on placebo. Reductions in BMI, waist circumference, systolic blood pressure and triglycerides were sustained at 5 mg as well as at the full dose. The authors wrote that reducing to 5 mg "might be a valuable alternative to discontinuation, although individual response may vary." A secondary analysis reported slightly different figures (22.4%, 17.0% and 10.1%) and estimated that the 5 mg dose retained about 70% of the prior loss, against essentially all of it on the full dose.

What it does and does not license

It licenses a conversation with a prescriber about stepping down to 5 mg after a successful course of tirzepatide, with the expectation of keeping most, not all, of the loss, and with a one-in-four chance of regaining enough to need the dose raised again within a year. It does not say anything about 2.5 mg, about 1 mg, about every-other-week dosing, or about semaglutide, for which no equivalent trial has been published. Our guide to stopping GLP-1s and keeping the weight off covers the trial in its own section and sets it beside the withdrawal trials.

The gap between 5 mg and a microdose

The distance is worth stating in numbers. Five milligrams is 200% of tirzepatide's starting dose. A microdose in the strict sense is a fraction of that starting dose, typically 20% to 50% of it, which is 10% to 25% of the dose SURMOUNT-MAINTAIN tested. The trial that is most often cited in support of microdosing tested a dose four to ten times larger than a microdose and found that it held about 70% of the loss, with a quarter of participants needing rescue. Whatever a true microdose does for maintenance, this trial did not measure it.

Microdosing is pitched, above all, as an alternative to stopping, so the case for it rests on how bad stopping is. The randomized withdrawal trials are unambiguous, and a 2026 meta-analysis put a rate on the regain.

The withdrawal trials

In the STEP 1 trial extension (Wilding and colleagues, Diabetes, Obesity and Metabolism, 2022, volume 24, pages 1553 to 1564), participants who had lost weight over 68 weeks on semaglutide 2.4 mg were followed for a year after treatment and lifestyle support ended: roughly two-thirds of the lost weight came back, and the cardiometabolic improvements largely reversed with it. In STEP 4 (Rubino and colleagues, JAMA, 2021, volume 325, pages 1414 to 1425), 803 people who had reached 2.4 mg after a 20-week run-in were randomized to continue or to switch to placebo for 48 weeks: those who continued lost a further 7.9% of body weight, those switched to placebo regained 6.9%. In SURMOUNT-4 (Aronne and colleagues, JAMA, 2024, volume 331, pages 38 to 48), 670 people who had lost a mean 21% over a 36-week tirzepatide lead-in were randomized to continue or to placebo for 52 weeks: continuation produced a further 5.5% loss, placebo a 14.0% regain, and a post hoc analysis in JAMA Internal Medicine found that 46% of the withdrawal arm still held at least a 10% loss from the study's start.

The 2026 meta-analysis: a rate, and a date

Sam West and colleagues at Oxford published a systematic review and meta-analysis in The BMJ on January 7, 2026 (DOI 10.1136/bmj-2025-085304) covering 37 studies and more than 9,000 participants, with a literature search through February 2025. As reported by the Nuffield Department of Primary Care Health Sciences and the UK Science Media Centre, regain after stopping any weight-management medication ran at about 0.4 kg a month, returning to baseline at roughly 1.7 years; for semaglutide and tirzepatide specifically, about 0.8 kg a month with a return to baseline near 1.5 years. Blood pressure and blood sugar were projected to return to pre-treatment levels in about 1.4 years. Regain after stopping medication was faster than after ending a behavioural programme, by roughly 0.3 kg a month, and behavioural support during drug treatment produced more loss during treatment but did not slow the regain afterwards. The authors noted wide individual variation and a small number of people who kept their loss.

Why this is the real argument for a lower dose

These data are the reason "less drug" is worth discussing at all. If stopping returns most of the loss in a year to eighteen months, then any dose that holds a meaningful share of it is worth having, and SURMOUNT-MAINTAIN shows that a labelled reduced dose can. The error in the microdosing argument is the next step: assuming that if 5 mg holds 70%, then a tenth of that holds something too. It might. The dose-response data say it would hold less, the records study says the lowest labelled dose produces modest results from scratch, and no one has measured it after a weight loss. Our guide to what happens when you stop goes through the biology of regain and the options that have been tested.

If microdosing is a response to stopping, it helps to know how often people stop and why, because the microdose programs are selling to exactly those people. The real-world persistence data are a mixed picture: most people stopped within a year in the earlier cohorts, and the trend is improving.

The large US cohort

Rodriguez and colleagues published "Discontinuation and Reinitiation of Dual-Labeled GLP-1 Receptor Agonists Among US Adults With Overweight or Obesity" in JAMA Network Open on January 31, 2025, using Truveta's electronic-records data on 125,474 adults who started a GLP-1 (liraglutide, semaglutide or tirzepatide). Overall, 53.6% had discontinued by one year and 72.2% by two years. Among people without type 2 diabetes, the group most likely to be using the drugs for weight alone, 64.8% had stopped at one year and 84.4% at two; among those with diabetes, 46.5% and 64.1%. Of those who stopped, 36.3% without diabetes and 47.3% with it restarted within a year. A Danish cohort reported in August 2025 found nearly half of adults who began semaglutide for obesity stopped within a year, with adults aged 18 to 30 nearly 50% more likely to stop than those aged 45 to 60, as Medscape reported.

The improving trend

Prime Therapeutics' pharmacy-claims analysis, reported by Medscape in 2026, found one-year persistence for weight-loss use nearly doubled from 33.2% in 2021 to 60.9% in the first half of 2024, with tirzepatide higher, at about 64% to 64.8%. Persistence and discontinuation are measured differently across these studies, so the figures are not interchangeable; the direction is consistent.

The clinician's version of the same numbers

At the Obesity Medicine Association's 2026 meeting, as Medscape reported, Craig Primack of Hims & Hers Health and Scottsdale Weight Loss Center argued the case for microdosing on side-effect grounds, citing that roughly 30% of patients have stopped by three months after starting and 40% to 50% within a year. Those are the same order of magnitude as the cohort data, and they are the practical reason the microdosing argument exists: a drug that half of patients abandon within a year is a drug with a tolerability problem, and a lower dose is the obvious first idea.

What the data say about why

The Truveta paper examined factors associated with stopping; the two that stand out in the coverage are gastrointestinal side effects and a weak early weight-loss response. Cost and coverage sit behind both in the real world, which is where the Evidation reasons come from. Each of those three reasons, side effects, cost and the wish to hold weight once lost, is examined on its own terms in the sections that follow, because the evidence for a lower dose differs sharply between them.

The strongest of the three arguments. Side effects are dose-dependent, the trials show it, and the one real-world low-dose study found fewer of them. The question is whether a microdose is the right tool, or whether the label's own slower ramp and step-down already do the job.

Side effects follow the dose

In STEP UP, gastrointestinal adverse events rose from 42.8% on placebo to 61.2% on 2.4 mg to 70.8% on 7.2 mg; dysesthesia from 0.3% to 6% to 22%. In the STEP trials behind the 2.4 mg dose, nausea affected roughly 40% to 44% of participants, diarrhea about 30%, vomiting around 25% and constipation about 24%, mostly during dose increases, as our semaglutide side-effects guide sets out. In the 2026 records study, sustained low-dose semaglutide was associated with fewer adverse events than sustained low-dose tirzepatide. The relationship is real: less drug, less nausea. It is also the relationship that gives less weight loss, which is the trade the microdosing pitch tends to leave out.

What the label already lets a clinician do

Both labels build in tolerability management without leaving the labelled range. Wegovy's allows holding any step for four extra weeks before increasing, and lowering 2.4 mg to 1.7 mg for up to four weeks. Zepbound's requires at least four weeks on each dose before a 2.5 mg increment and offers three maintenance doses, with the label directing clinicians to choose on tolerability as well as response. The 2026 Zepbound label added 7.5 and 12.5 mg strengths, which make the increments smaller. A prescriber who wants to reduce side effects therefore has several moves before anything unlabelled: slow the ramp, stay a step longer, step down a rung, choose the lowest maintenance dose. The academic-clinic data in a 2025 real-world study of titration (published in PMC) show that in practice many patients never reach the top of the ladder at all, and are managed at intermediate doses.

Where microdosing fits, honestly

For a patient who cannot tolerate even 0.25 mg of semaglutide or 2.5 mg of tirzepatide, a smaller dose is a reasonable clinical idea, and it is the situation Primack described at OMA 2026. Nothing about it is tested. The OMA working group's statement notes that no data exist on the long-term health impact of microdosed obesity medications, and Dushay's STAT essay points out that none of the compounded products used for it have been studied for safety, efficacy or weight regain. So the honest framing is: a slow ramp and labelled step-downs first, because they are tested; an off-label fraction of the starting dose only as a prescriber's judgement call for someone who cannot tolerate the floor, with the expectation of modest weight loss and no outcome data.

A note on the oral GLP-1s

Two pills now offer a different route to tolerability. The Wegovy pill (oral semaglutide 25 mg) and Foundayo (orforglipron) are started low and stepped up like the injections, and in their trials gastrointestinal effects were, as for every GLP-1, mostly early and mostly during escalation. For someone whose problem with an injectable is the injection, a pill is a tested alternative; for someone whose problem is nausea, it is the same class with the same profile. Our pills vs injections guide sets the two side by side.

The clearest statement of both sides came at the Obesity Medicine Association's 2026 meeting, where microdosing was debated as one of four "weight-loss controversies," as Medscape reported.

For: a tool for the patients who quit

Craig Primack, of Hims & Hers Health and Scottsdale Weight Loss Center, argued that microdosing "can be an effective strategy in various situations," with side-effect management as the central case: about 30% of patients have stopped by three months and 40% to 50% within a year, and a dose a patient can tolerate is better than a dose they abandon. The argument is practical rather than evidential, and it is the one most clinicians who prescribe lower doses would recognise: the alternative to a small dose is often no dose.

Against: an intervention nobody can define

Leslie Golden, founder and CEO of Weight In Gold Wellness and chair of the OMA's Advocacy Committee, replied that "an undefined intervention can't be consistently put into practice." Her objection is the one the working-group statement makes in writing: microdosing is "a marketing term," there is "little to no peer-reviewed evidence" for the safety or effectiveness of microdosed regimens for obesity, and "no data exists on the long-term health impact." A practice that cannot be specified cannot be studied, and a practice that cannot be studied cannot be recommended.

The academic voices

Dushay's STAT essay of May 29, 2026 adds three specific concerns: that the compounded products used for microdosing have not been studied for safety, efficacy or weight regain; that compounded vial "units" do not map onto the branded medications' milligrams; and that some clinics draw less than a full single-use Zepbound vial, a practice that has not been studied and carries contamination risk. Velazquez's Cedars-Sinai page states that there is no evidence that low-dose, intermittent GLP-1 use gives meaningful or lasting weight loss or broader health benefits, while allowing that lower doses "may help with weight maintenance after initial treatment," which is the SURMOUNT-MAINTAIN point. A September 1, 2026 review summarised by News-Medical called the direct evidence "extremely limited." Science News, in a piece by Jamie Ducharme ("GLP-1 microdosers are chasing longevity"), quoted a bariatric specialist saying there is no rigorous scientific data to support it.

Where the two sides actually agree

Read closely, Primack and Golden are not disagreeing about evidence. Both accept there is none. They disagree about what a clinician should do in its absence: use judgement with an individual patient who is otherwise going to quit, or decline to practise an intervention that has no definition. That is a genuine clinical disagreement, and it is the right frame for a patient too. A lower dose chosen with a prescriber for a specific tolerability problem is one thing; a "microdosing program" sold to the public as a maintenance or longevity strategy is another, and the societies' objection is aimed at the second.

The newest and least-evidenced branch of the trend is microdosing by people who are not trying to lose weight at all. The pitch is that GLP-1s have benefits beyond weight, which is true at full doses in people with obesity or diabetes, and that a small dose might capture them without the weight loss, which no one has shown.

Where the idea comes from

Semaglutide's cardiovascular-outcomes trial (SELECT) and the kidney, sleep-apnea and heart-failure trials that followed established benefits at full doses in people who qualified for treatment. Early company-reported findings in 2026 went further: MIT Technology Review reported on October 9, 2026 that Eli Lilly and Novo Nordisk have presented analyses suggesting the drugs may lower biological age by about two to three years in people with overweight or diabetes, and noted in the same sentence that those results come from the drugmakers themselves. None of this work studied small doses, and none of it studied people of normal weight.

The one trial

The only registered study of microdosing for longevity-type outcomes is NCT07092605, "Effectiveness of Microdosed GLP-1 in Improving Health, Quality of Life, and Longevity Measures." Its consent form names the sponsor as AgelessRx, a telehealth company that sells microdoses, and describes a decentralised study run over telemedicine of microdosed semaglutide. AgelessRx's own page said in 2026 that enrolment had concluded and the study was awaiting the last participants to finish before analysis. No results have been published. The sponsor's commercial interest is the obvious caveat; the design, a single-arm telemedicine study without a placebo group, is the larger one, because a study like that can report changes but cannot attribute them to the drug.

The observational cohort

A prospective cohort registered as NCT07588984 (FLOURISH and THRIVE), with an estimated 2,310 enrollees, lists microdosing among its keywords and may eventually describe how people who microdose fare against those who do not. It has reported nothing yet.

What a person of normal weight is actually buying

For someone without obesity, a GLP-1 at any dose is off-label, untested for the purpose, and carries the class warnings: the boxed warning on thyroid C-cell tumours seen in rodents, pancreatitis, gallbladder disease and, in the labels, a caution about use in people without the conditions studied. The 2026 joint advisory from four societies, discussed below, warns that intermittent use and weight cycling may worsen muscle and bone loss. Someone taking a microdose for longevity is, on current evidence, accepting a drug's risks for a benefit that has not been measured at that dose in that population. That may change if the registered studies report; as of October 2026 they have not.

The approved injectables come in pens that deliver fixed doses. You cannot easily take a tenth of a Wegovy pen or a quarter of a Zepbound KwikPen, and you are not supposed to try. Compounded semaglutide and tirzepatide come in multi-dose vials, drawn into an insulin syringe, and a vial can be drawn to any volume. That single difference is why Evidation found about half of microdosers using compounded versions, why the microdose programs sell compounded products, and why the FDA's adverse-event reports for compounded GLP-1s are, to a large degree, reports of dosing errors.

The FDA's dosing-error alert

On July 26, 2024 the FDA issued an alert on dosing errors with compounded semaglutide, now folded into its page on its concerns with unapproved GLP-1 drugs used for weight loss, which covers both compounded semaglutide and tirzepatide and confirms that some reported events required hospitalisation. As CBS News and Renal and Urology News reported from the alert, most reports involved patients drawing too much from a multiple-dose vial, with some taking five to twenty times the intended dose. Instructions written in "units" rather than milligrams or millilitres contributed: in one case a 5-unit dose was given as 50 units. Prescribers made errors too, including a 0.25 mg dose, 5 units, prescribed as 25 units. Many affected patients had little experience with self-injection and did not know how to draw medication from a vial. The reported harms included nausea, vomiting, abdominal pain, fainting, headache, dehydration, acute pancreatitis and gallstones. The FDA's advice was that providers supply an appropriately sized syringe and counsel patients on measuring the dose, and that adverse events be reported to MedWatch.

Why microdosing makes the error more likely

A microdose is, by definition, a small volume. The smaller the intended volume, the larger the proportional error from a misread syringe line or a units-for-milligrams confusion, and the easier it is to take several times the intended dose. The FDA's cases are mostly people who meant to take a starting dose and took far more. A patient measuring a fraction of a starting dose is doing the same task with a smaller target. Dushay's essay adds the second problem: compounded vial "units" do not map onto branded milligrams, so a patient who moves between products or reads a social-media chart calibrated to a different concentration is working from the wrong scale.

The vial-splitting problem

Zepbound is also sold in single-dose vials, meant to be drawn once and discarded. Dushay reports that some clinics draw less than a full single-use vial, keep the remainder and use it later, a practice that has not been studied and that carries contamination risk, because a single-dose vial has no preservative and is not designed to be re-entered. Our guide to 503A and 503B pharmacies and the 28-day rule explains why compounded multi-dose vials carry a beyond-use date and why stretching a vial past it is not a saving.

What this means for anyone considering it

If a lower dose is prescribed, the safe version is one where the dose is written in milligrams, the concentration of the vial is on the label, the syringe is the right size for the volume, the first draw is demonstrated by someone who knows how, and nobody is re-entering a single-dose vial. Every one of those is a question you can ask a program before paying, and the FDA's own guidance is the standard to hold it to.

Microdosing sits inside the larger story of compounded GLP-1s, which has moved fast in 2026. The short version: compounding these drugs in bulk is over, patient-specific compounding continues under conditions, and the FDA has spent the year writing to the telehealth companies that market the products.

The shortage is over, so the exemption is over

Compounding pharmacies were allowed to make copies of semaglutide and tirzepatide while the FDA listed them as in shortage. The FDA declared the tirzepatide shortage resolved on December 19, 2024 and the semaglutide shortage on February 21, 2025. Since then a compounded GLP-1 is legal only as a patient-specific prescription from a state-licensed 503A pharmacy, for a documented clinical need that the approved product does not meet. "A lower dose than is marketed" is one of the clinical needs some prescribers document, which is the legal foothold for microdose programs; whether a given prescription meets that standard is a matter between the prescriber, the pharmacy and their state boards, and the FDA has not blessed it.

The April 2026 proposal

On April 30, 2026 the FDA proposed to exclude semaglutide, tirzepatide and liraglutide from the 503B bulks list, saying that after careful review it "did not identify sufficient clinical need for outsourcing facilities to compound these drugs from bulk substances" (FDA statement; analysis by Orrick, May 2026). The proposal targets 503B outsourcing facilities, the larger-batch manufacturers; 503A pharmacies, which fill individual prescriptions, are not its subject. The comment period closed on July 30, 2026 after a 30-day extension, and as of the reporting we have seen no final rule had been issued. If it is finalised, the supply of compounded GLP-1s narrows to patient-specific 503A compounding, which is also where microdosing lives.

The warning letters

The FDA's enforcement has focused on marketing. A round of 30 warning letters to telehealth firms, dated February 20, 2026 and made public on March 3, cited claims implying that compounded products are the same as the FDA-approved drugs and the practice of branding compounded products under the telehealth firm's own name (Foley & Lardner analysis, March 2026). A further 25 letters went out in June 2026 (Telehealth.org), following roughly 80 warning letters and 40 untitled letters in September 2025. The recurring theme is the "misleading net impression" that a compounded product is approved or equivalent. A microdose program that describes its product as "low-dose Wegovy" is inside that theme.

The Hims & Hers case

On February 9, 2026 Novo Nordisk sued Hims & Hers in the District of Delaware over compounded semaglutide sold on its platform (CNBC, February 9, 2026). Hims had defended compounding as access to "personalised care," the framing under which lower-than-labelled doses were offered. On March 9, 2026 Novo Nordisk dismissed the claims without prejudice after Hims agreed to stop promoting compounded GLP-1 products and to offer FDA-approved Ozempic and Wegovy at Novo's self-pay prices (HMP Global, March 2026). The largest telehealth seller of "personalised" compounded semaglutide, in other words, has moved to the branded products.

What is unchanged

None of this makes a lower dose illegal. A licensed prescriber can prescribe off-label, a 503A pharmacy can fill a patient-specific prescription, and a patient can take it. What has changed is the marketing environment, the supply of bulk-compounded product and the legal posture of the largest sellers. For an in-depth treatment of the framework, see our compounded vs brand-name semaglutide guide.

The commercial side is where the trend became a category. We describe the programs from their own published terms and dated reporting, and we quote prices only where we have a dated source.

Noom

Noom launched its microdose program in May 2025. Pharmaceutical Commerce reported on May 21, 2025 that it sold compounded semaglutide (the headline's "Wegovy" was a misnomer the article itself corrected) at $149 for the first month, against $349 for a vial of branded Wegovy or Zepbound through Noom, on a plan that started at 0.25 mg and rose toward 2.4 mg over 20 weeks, which is the labelled ladder, slowed. Reuters, in August 2025, reported a revised offer: $119 for the first month, then $199 a month, for a maximum dose of 0.6 mg. Our own capture of Noom's published pricing, on the Noom subscription cost page, recorded the Microdose GLP-1 plan at $99 for an initial three-week subscription plus four weeks of medication, then $199 a month with medication included, described by Noom as an entry-level option "using a lower dose of compounded semaglutide." Third-party reviews in 2026 reported further changes to the introductory price; we have not verified those and do not quote them. The pattern across every version is the same: a lower dose, a lower first-month price, then a monthly price in the $179 to $199 range.

The wider market

Forbes Health's 2026 survey of microdosing costs (Forbes Health) put compounded microdose plans generally at $149 to $199 a month, with month-to-month plans "around $300 or more" and the lowest prices often requiring payment for several months or a year upfront. Forbes also warned that compounded products carry increased risk of "dosing errors, unknown potency, unapproved ingredients, contamination." Those figures are reported ranges across programs we have not individually verified.

What is not in the microdose market

None of the nine providers in our verified GLP-1 price index, checked October 16, 2026, lists a microdose plan. Their published prices are for labelled-dose plans, and several price them flat at every dose: welludr at $49 a month for semaglutide and $89 for tirzepatide (12-month plan), altRx at $89 and $149, DirectMeds at $147 for either, TrimRx at $149 for semaglutide. That fact is the foundation of the cost section that follows, because it means the cheapest way to take semaglutide at any dose on our index is a full labelled dose at a flat-price provider, not a microdose at a microdose program.

The branded direct-pay programs

The manufacturers' own cash prices have also fallen, which changes the comparison the microdose programs were built on. Lilly's Foundayo pill lists from $149 a month at its lowest dose, as reported from Lilly material, and the Wegovy pill is $149 for the first month then $299 at ro, the one telehealth price we have verified. Our Wegovy pill price guide and lost-coverage guide carry the current figures and their sources. A microdose program at $199 a month is now competing with FDA-approved products at a starting dose for a similar or lower price.

Cost was the second most common reason for microdosing in the Evidation data, at 31.9%. The arithmetic deserves a plain answer, and at the providers we track the answer is mostly no.

Flat pricing makes the dose irrelevant

Four of the nine providers in our verified index charge the same price at every dose: welludr ($49 semaglutide, $89 tirzepatide, 12-month plan), altRx ($89 and $149), DirectMeds ($147 for either) and TrimRx ($149 semaglutide). At those providers a 0.25 mg dose costs exactly what a 2.4 mg dose costs. There is no saving in taking less, because you are paying for the plan, not the milligrams. The same is true of embody ($49 and $99, month to month) and Medvi ($99 and $166 promotional, all-inclusive), whose prices do not vary by dose either. Every price here is from the cheapest GLP-1 index, verified October 16, 2026.

The microdose program costs more than a full dose elsewhere

Put the two side by side. Noom's microdose plan, at its $199 monthly rate for up to 0.6 mg of compounded semaglutide, costs four times welludr's $49 for a full labelled dose of compounded semaglutide, more than twice embody's $49 and altRx's $89, and twice Medvi's $99 all-inclusive price. Over a year that is $2,388 against $588, $588, $1,068 or $1,188. Someone choosing a microdose program to save money would, on our index, pay more for less medication than at any of those providers. The comparison is not quite like for like, since Noom's plan includes its coaching app, but the Evidation respondents said cost, not coaching, was their reason.

Where a lower dose does save: vial stretching, and why it is a false economy

The saving people imagine comes from making a vial last longer: a vial sized for a month at 2.4 mg lasts several months at a tenth of that. Three things stand against it. The 28-day rule: an opened multi-dose vial of a compounded sterile product is generally assigned a 28-day use period, after which it is to be discarded, for sterility reasons rather than potency ones, as our 503A/503B guide explains; stretching past it is not a saving, it is an expired product. The dosing-error risk: a smaller draw from the same vial is the mechanism in the FDA's reports. And the provider terms: a monthly plan ships monthly, so a patient taking less does not receive less or pay less; the vial arrives on schedule regardless.

Brand-name pens: no fractional option

A Wegovy or Zepbound pen is a fixed dose. The branded route offers no microdose at all, and its prices are set by dose tier at the manufacturers' direct-pay programs and by pen at the telehealth providers (Zepbound at $1,249 a month at altRx and $1,599 at welludr, cash, from our index). The only way to pay less on a branded injectable is a lower labelled tier where the program prices by tier, insurance, or the Medicare GLP-1 Bridge's flat $50 copay for eligible beneficiaries, which does not vary by dose, as our Medicare guide sets out.

The honest cost answer

If the goal is the cheapest safe way to take a GLP-1, our index says it is a full labelled dose of compounded semaglutide at a flat-price provider, from $49 a month on a 12-month plan or, during embody's pre-Black Friday sale, $49 month to month, with the clinician review and delivery included, from a named 503A pharmacy. If the goal is the cheapest FDA-approved product, it is a pill or pen at the manufacturer's lowest tier, with insurance or Medicare if either applies. In neither case is a microdose the cheaper route. Our month 1 vs year one guide does the full-year arithmetic for every provider.

One source of evidence about low doses in practice is not a trial at all: the reviews we transcribe from each provider's Trustpilot profile and group into themes on our review pages. Dosing comes up in them, and the pattern is instructive for anyone thinking about a lower dose through a telehealth program.

Three recurring dose complaints

On the ro review, a small group of reviewers describe dosing decided before any conversation: being restarted at the lowest dose after a higher one elsewhere, choosing a medication and being charged before physician input, and a smaller starting dose than friends received. On the Medvi review, reviewers transferring from another provider describe being sent a starting dose instead of their current one, with a five-star reviewer noting that the doctor spotted and fixed the same mistake. On the embody review, the most serious complaints in the stored set are a prescription that arrived at roughly triple the expected dose, a next order placed at a different dose than discussed, and a formulation mixed with another ingredient unless the patient asked otherwise. Each of these is a handful of reviews among dozens, counted on the review page from the stored set, and each provider's praise themes are larger than its complaint themes.

Why it matters for a lower dose

Every one of those complaints is about the dose that arrived not matching the dose that was discussed. That is the ordinary failure mode of mail-order prescribing, and a lower-than-labelled dose adds two more places for it to happen: the prescription itself, where a fraction of a milligram has to be written and filled correctly, and the vial, where it has to be drawn correctly. A program whose reviews already describe dose mismatches at labelled doses is not a program to trust with an unlabelled one. The counts behind each theme, with the reviewers' own words, are on the review pages, and the methodology page explains how they are collected.

What the praise themes say

The other side of the same reviews is relevant too. The most-mentioned praise theme at embody, Medvi and Found is clinicians who take time, explain dosing and adjust the plan when asked; at welludr and Eden it is support staff who fix a problem by name. Those are the providers where a conversation about a lower dose is most likely to be a conversation. The provider-choosing guide sets out how to check clinician involvement before you sign up.

One cost of microdosing is rarely discussed because it is not about the drug but about the pattern of use. A dose that produces a little loss, then a little regain, then a restart, is weight cycling, and the professional societies have now said in writing what that does to muscle and bone.

The advisory

In April 2026 the American College of Lifestyle Medicine, the American Society for Nutrition, the Obesity Medicine Association and The Obesity Society published a joint advisory, "Nutritional priorities to support GLP-1 therapy for obesity" (Mozaffarian and colleagues, Obesity, first published online April 23, 2026). Two of its warnings bear directly on low-dose and intermittent use. First, that intermittent GLP-1 use, along with weight regain, may worsen muscle and bone loss. Second, that protein intake alone is likely insufficient to preserve muscle without resistance training. The advisory's nutrition guidance, which our what to eat on a GLP-1 and exercise on a GLP-1 guides summarise, assumes a course of treatment at an effective dose with deliberate muscle protection, not a sawtooth.

Why cycling matters more than the dose

When weight is lost, a share of it is lean tissue; when it is regained, the regain is disproportionately fat unless resistance training and protein drive it back toward muscle. Each cycle therefore tends to shift body composition the wrong way. The withdrawal trials measured the weight; the advisory's concern is what the weight is made of. A microdose that holds weight imperfectly, or an every-other-week schedule that lets appetite return between doses, is a recipe for small cycles, which is the pattern the advisory names.

What this adds to the decision

Whatever dose you and your prescriber settle on, the advisory's two instructions are the same: enough protein, spread across the day, and resistance training two to three times a week. For someone considering a lower dose specifically to hold weight, the body-composition argument favours a dose that actually holds it, which on current evidence means a labelled maintenance dose rather than a fraction of one.

This section is a framework for a conversation with a prescriber, not a recommendation. The evidence supports some lower-dose situations and gives no support to others.

Situations where the evidence supports a lower labelled dose

After a successful course of tirzepatide, to hold the loss. SURMOUNT-MAINTAIN is direct evidence that stepping down to 5 mg kept most of the loss for a year, with a quarter of participants needing the dose raised again. That is a labelled dose, tested for this purpose, and the strongest case in this guide for taking less.

During escalation, when a step is not tolerated. Both labels allow a slower ramp and a temporary step down. The tested move is to stay longer on the current rung or drop one, not to drop below the ladder.

At maintenance, on semaglutide, choosing 1.7 mg over 2.4 mg. Both are labelled maintenance doses; the label leaves the choice to response and tolerability. There is no semaglutide equivalent of SURMOUNT-MAINTAIN yet, so the lower maintenance dose is labelled but not trial-tested for holding weight specifically.

Situations where a lower dose is a judgement call with no data

Someone who cannot tolerate the starting dose. A fraction of 0.25 mg or 2.5 mg is the situation Primack described at OMA 2026: a reasonable clinical idea for a specific patient, untested, with modest expected effect. It belongs to a prescriber who knows the patient, with the dose written in milligrams and the measurement demonstrated.

Someone whose cost constraint is real. See the cost section: at flat-price providers a lower dose saves nothing, and a microdose program costs more than a full dose elsewhere. The cheaper answer is usually a different provider, not a smaller dose.

Situations where microdosing has no support and real risk

Someone who has never been treated and expects trial results. The records study and the dose-response trials say the lowest doses produce a few percent of body weight over a year. A person with obesity starting at a microdose is choosing the least effective tested option and then going below it.

Someone of normal weight taking it for longevity or appetite. Off-label, untested at the dose, untested in the population, with the class risks intact and the registered study unreported.

Anyone measuring their own fraction from a vial without training. The FDA's dosing-error reports are this scenario. More than a third of the Evidation microdosers were deciding doses without a provider.

Anyone with a history of an eating disorder. Clinicians writing in 2026 on the overlap between GLP-1 use and eating disorders (STAT, February 19, 2026) have warned that appetite-suppressing drugs used outside a treatment plan can feed restriction; a self-managed microdose is the least supervised version of that.

Anyone pregnant, planning pregnancy or breastfeeding. The labels advise against use; the dose does not change that.

The question that sorts most cases

Ask what the lower dose is for. If the answer is "to hold what I lost," the tested answer is a labelled maintenance dose. If it is "to tolerate the ramp," the tested answer is a slower ramp. If it is "to save money," the arithmetic points elsewhere. If it is "for longevity," there is nothing to point to yet. The cases left over, the ones with no tested answer, are the ones to take to a prescriber rather than a program.

A lower dose is a clinical decision, and a good one is made with information the prescriber has and you may not. These are the questions that turn a vague wish to "microdose" into a specific, safer plan.

Questions about the plan

Which of the three things do we mean? A labelled lower dose, a dose below the starting dose, or a labelled dose less often. Each has a different evidence base, and naming it first avoids talking past each other.

What is the dose in milligrams, and what is the concentration of what I will receive? Insist on both numbers. The FDA's reports are largely about "units" standing in for milligrams and patients drawing from vials whose concentration they did not know.

What am I trying to achieve, and how will we know? A target weight to hold, a side effect to resolve, a timeline to review. SURMOUNT-MAINTAIN's rescue criterion, regaining half of what was lost, is a usable review point for a maintenance dose.

Have we used the label's own tools first? A slower ramp, a longer hold on a step, a step down to the previous rung, the lowest labelled maintenance dose. These are tested; the question is whether they have been tried.

Questions about the product

Is this a brand-name pen or a compounded vial, and which pharmacy makes it? A pen cannot be microdosed; a vial can, which is where the errors happen. If compounded, the pharmacy's name and state licence, and whether it is 503A, are the questions our pharmacy guide sets out.

Will someone show me the first draw? The FDA's advice to providers is a correctly sized syringe and counselling on measuring the dose. A program that ships a vial and a chart has not done that.

What is the beyond-use date, and how many doses does this vial actually contain at my dose? So that the plan does not quietly become vial-stretching past 28 days.

Questions about cost

Does the price change with the dose? At the providers we track it does not. If the proposal is to pay the same for less medication, the saving is imaginary.

What is the full-dose price at a flat-price provider? Bring the number: $49 to $99 a month for compounded semaglutide at verified prices, with the review and delivery included. A lower dose should be chosen for a clinical reason, and the cost reason should be checked against that figure.

What a good answer sounds like

A prescriber who says "let's stay on this step for another month" or "let's drop you to 5 mg and review in twelve weeks" is using the label and the trial evidence. A prescriber who says "we'll try a smaller dose because you can't tolerate the starting dose, here is exactly how much in milligrams and here is how to draw it" is using judgement in the absence of evidence, and saying so. A program that says "microdosing preserves your results with fewer side effects at a lower cost" is making three claims, none of which the evidence supports as stated.

The FDA's warning letters, the societies' statements and the dosing-error reports together describe the program to walk away from. These are the signs, each tied to the source that makes it a problem.

It calls the product "low-dose Wegovy" or "microdose Ozempic." A compounded product is not the branded drug, and implying equivalence is the central charge in the FDA's 2026 warning letters. The honest description is "compounded semaglutide" with the pharmacy named.

The dose is given only in units. The FDA's dosing-error cases turn on units standing in for milligrams. A legitimate prescription states the dose in milligrams and the vial states its concentration.

No clinician reviews you, or the review is a form with no possibility of being declined. Compounded GLP-1s are legal only as patient-specific prescriptions for a documented clinical need. A program that prescribes to everyone is not documenting a need.

It claims microdosing preserves weight loss, prevents regain or extends life. No trial supports the first two at a microdose; the third is a registered study with no results. The OMA working group's wording is "little to no peer-reviewed evidence."

It sells a split single-dose vial, or tells you to make a vial last beyond its beyond-use date. Both are the practices Dushay and the sterile-compounding standards warn against.

It will not name the pharmacy. The pharmacy is the part of the operation that decides what is in the vial. A provider that hides it cannot be checked.

It prices the plan to look cheaper than a full dose, when it is not. Compare the monthly figure with the flat-price providers in our index. A $199 microdose plan is not a budget option next to a $49 to $99 full-dose plan.

It markets to people of normal weight. The approved indications are obesity, or overweight with a weight-related condition. A program advertising GLP-1 microdoses for "wellness," "appetite" or "longevity" to people outside those criteria is selling an off-label use with no supporting data.

The positive version of this list is the one in our guide to choosing a GLP-1 provider: a licensed clinician who can decline, a named state-licensed pharmacy, published prices with the regular rate beside any promotion, and a public review record you can read yourself.

The evidence base for microdosing is thin because the question is new, not because it is unanswerable. Several things could change the picture over the next two years.

A reported microdosing trial

AgelessRx's NCT07092605 has finished enrolling and will, at some point, report. A single-arm telemedicine study cannot settle efficacy, but it will be the first published data on people deliberately taking semaglutide below the starting dose, and the adverse-event and adherence findings will be informative whatever the headline. The FLOURISH and THRIVE cohort (NCT07588984) may add observational comparisons.

Maintenance-dose trials for semaglutide

SURMOUNT-MAINTAIN answered the reduced-dose question for tirzepatide. A semaglutide equivalent, testing 1.7 mg or lower against 2.4 mg and placebo after weight loss, would answer it for the more widely used drug. None has reported; the drug's own development is heading upward, to 7.2 mg.

The FDA's 503B rule

If the April 30, 2026 proposal to exclude semaglutide, tirzepatide and liraglutide from the 503B bulks list is finalised, bulk-compounded supply ends and every compounded GLP-1 becomes a patient-specific 503A prescription. That narrows the microdose market to prescriptions a clinician will document, and it may change what the programs can sell.

The next drugs

The FDA was expected to decide on CagriSema, Novo Nordisk's semaglutide and cagrilintide combination, by the end of 2026, as Time reported on October 1, 2026, and Lilly has said it plans to file retatrutide in the first quarter of 2027 (Lilly). Every one of these is a more potent agent at a labelled dose. The industry's answer to tolerability has been slower ramps, more strengths and new molecules, not smaller doses of the old ones.

The prices

The microdose programs were priced against $349 vials of branded medication in 2025. In 2026 the branded pills start at a reported $149 a month and our verified compounded index starts at $49. If branded prices keep falling, the microdose program's cost argument, already weak on our index, disappears.

What we will update

This guide carries the date of its last revision. When any of the studies above reports, when the FDA finalises or withdraws the 503B rule, or when a provider in our index publishes a microdose plan, the relevant section will be revised with the new source and date.

The evidence in this guide is easier to use as three situations than as a list of trials. None of these is a recommendation; each is the evidence applied to a common case, with the decision left where it belongs, with a prescriber.

The person who lost the weight and wants to take less

Eighteen months on tirzepatide, 20% down, tolerating 10 mg, tired of the cost and the weekly injection. The tested answer is SURMOUNT-MAINTAIN: a step down to 5 mg kept about 70% of the loss over a year, with a one-in-four chance of needing the dose raised again. The untested answer is a fraction of 2.5 mg from a compounded vial, which no trial has studied after a weight loss and which the dose-response data suggest would hold less. The cost answer is that at a flat-price provider the step to 5 mg saves nothing on the bill, and that the cheapest full-dose plan on our index is $89 a month for compounded tirzepatide, so the budget question is really about which provider, not which dose. A review point, borrowed from the trial, is regaining half of what was lost.

The person who cannot get past the first month

Two weeks on 0.25 mg of semaglutide, nausea most days, ready to quit. The label's answer comes first: hold the dose for another four weeks before stepping up, eat smaller and blander meals, and treat the escalation as a schedule to be slowed rather than kept. The side-effect guide covers what helps at each step. If the starting dose itself is intolerable, a smaller dose is the judgement call Primack described, with the expectation of modest weight loss and the measurement risks of a vial; that is a conversation with the prescriber, with the dose in milligrams, not a program's chart. The alternative with a tested profile is a different molecule: tirzepatide's GI rates in its trials were lower than semaglutide's at comparable effect, and the pills offer a different route. None of those is a microdose, and all of them have data.

The person of normal weight who wants the "benefits" at a small dose

Not overweight, drawn by reports of lower biological age, reduced alcohol interest or heart protection, considering a microdose from an online program. Every benefit on that list was measured at full doses in people with obesity or diabetes; the biological-age analyses are company-reported; the one microdosing-for-longevity study is sponsored by a seller and unreported. The class warnings apply at any dose, the joint advisory's concern about muscle and bone applies to intermittent use, and the FDA's dosing-error reports apply to anyone drawing small volumes from a vial. On current evidence this person is accepting a drug's risks for a benefit nobody has measured at that dose in that population. The honest answer is to wait for the registered studies to report.

The common thread

In all three cases the lower-dose idea is reasonable and the microdose version of it is the untested one. The labelled tools, a slower ramp, a step down, a maintenance dose, a different molecule, a flat-price provider, cover most of what people are actually trying to achieve. What is left over is a small set of individual clinical decisions, which is exactly where the Obesity Medicine Association's debate left it too.

Does microdosing a GLP-1 work for weight loss? Not in any way that has been measured. No trial has tested a dose below the starting dose. People who stayed on the lowest approved doses for a year lost about 2.2% (semaglutide) or 5.5% (tirzepatide) of body weight in a 2026 records study, against roughly 15% to 21% at full doses in the trials.

Does microdosing keep weight off after you stop losing? Untested at a microdose. The tested lower dose is 5 mg of tirzepatide, which held about 70% of the loss in SURMOUNT-MAINTAIN, with a quarter of participants needing the dose raised within a year.

Does a microdose have fewer side effects? Side effects fall with dose in every trial, so a smaller dose should produce fewer. It also produces less weight loss, and the labels already allow a slower ramp and a step down without going below the starting dose.

Is microdosing cheaper? Usually not. Four of the nine providers in our verified index charge the same at every dose, so a lower dose costs the same as a full one, and the best-known microdose program's $199 monthly rate is more than a full dose at welludr ($49), embody ($49), altRx ($89) or Medvi ($99).

Is microdosing legal? A licensed prescriber can prescribe off-label and a 503A pharmacy can fill a patient-specific prescription for a documented clinical need. Marketing a compounded product as equivalent to the branded drug is what the FDA's 2026 warning letters target.

Can you microdose a Wegovy or Zepbound pen? No. Pens deliver fixed doses. Microdosing is done with compounded vials, which is why about half of microdosers in the Evidation survey were using compounded products.

Why do doctors warn about it? Three reasons in the statements: no evidence, no definition, and the FDA's reports of dosing errors from patients drawing small volumes from vials, including cases of five to twenty times the intended dose.

What about microdosing for longevity? The only registered study, sponsored by a company that sells microdoses, has not reported. The company-reported "biological age" analyses from Lilly and Novo Nordisk used full doses in people with obesity or diabetes.

What is the tested alternative to microdosing? For maintenance, a labelled reduced dose (5 mg tirzepatide in SURMOUNT-MAINTAIN). For tolerability, the label's slower ramp and step-downs. For cost, a flat-price provider at a full dose.

Where should I start if I want to take less? With the prescriber who manages your treatment, with the dose in milligrams, and with the questions in the section above.

Every clinical figure in this guide is taken from the trial report, label, meta-analysis, registered trial or professional statement linked next to it, with its publication date. Where a figure reached us through press coverage, conference reporting or a company release rather than a peer-reviewed paper, the outlet and date are given and the figure is described as reported. Every provider price is from our verified GLP-1 price index, checked on each provider's site on October 16, 2026, or from the dated program page we link. We did not estimate any number, and we did not publish any dose below the labelled starting doses, because none has been tested.

Primary sources, in order of appearance

• Obesity Medicine Association Access and Policy Working Group, Microdosing of obesity medications: a perspective statement (2026).
• Dushay J., STAT, May 29, 2026; Cedars-Sinai, Can microdosing GLP-1s promote health and weight loss? (2026).
• Evidation survey: Medscape, January 27, 2026; ISPOR 2026 poster.
• Gallup National Health and Well-Being Index, as reported July 10, 2026; EY-Parthenon research, CBS News, October 22, 2026.
• Wegovy dosing: prescribing information via Drugs.com; Wegovy HD approval: Novo Nordisk, March 19, 2026; Zepbound: FDA label, 2026.
• O'Neil et al., The Lancet, 2018 (semaglutide dose-ranging); Jastreboff et al., NEJM, 2022 (SURMOUNT-1); STEP UP, The Lancet Diabetes & Endocrinology, November 2025.
• Murugadoss et al., Biology Methods and Protocols, August 31, 2026; Healthline, September 18, 2026.
• SURMOUNT-MAINTAIN, The Lancet, May 2026; ACC summary, May 18, 2026.
• Wilding et al., Diabetes, Obesity and Metabolism, 2022 (STEP 1 extension); Rubino et al., JAMA, 2021 (STEP 4); Aronne et al., JAMA, 2024 (SURMOUNT-4); West et al., The BMJ, January 7, 2026.
• Rodriguez et al., JAMA Network Open, January 31, 2025; Prime Therapeutics persistence data via Medscape, 2026.
• Medscape coverage of Obesity Medicine 2026.
• FDA, FDA's concerns with unapproved GLP-1 drugs used for weight loss; dosing-error alert of July 26, 2024 as reported by CBS News.
• FDA 503B proposal, April 30, 2026 (Orrick analysis); warning letters (Foley & Lardner, March 2026); Novo Nordisk v. Hims & Hers (CNBC, February 9, 2026).
• Noom microdose program: Pharmaceutical Commerce, May 21, 2025; Reuters via PharmaLive, August 2025; Forbes Health, 2026.
• Mozaffarian et al., Obesity, April 23, 2026 (joint advisory); MIT Technology Review, October 9, 2026; Science News; NCT07092605; Time, October 1, 2026.

Where to go next

If your question is really about stopping, our guide to stopping GLP-1s and keeping the weight off is the long version of the regain and maintenance evidence. If it is about side effects, the semaglutide side-effects guide covers what helps at each step. If it is about cost, the cheapest GLP-1 index and the month 1 vs year one guide carry every verified price. If it is about whether a compounded product is safe, the compounded vs brand-name guide and the 503A vs 503B guide explain how to check the pharmacy. And if you are starting from the beginning, the complete guide to GLP-1 weight loss online is the place to begin. This article is general information, not medical advice; a licensed clinician decides whether any dose of any GLP-1 is appropriate for you.