Research
STEP 4: the semaglutide withdrawal trial
STEP 4 (NCT03548987) is the semaglutide maintenance/withdrawal trial: after a 20-week run-in, participants who continued semaglutide kept losing or he
STEP 4 (NCT03548987) is the semaglutide maintenance/withdrawal trial: after a 20-week run-in, participants who continued semaglutide kept losing or held weight, while those switched to placebo regained a large share over the next 48 weeks. It is the core evidence that stopping leads to regain.
What the trials establish
STEP 4 was designed specifically to answer the maintenance question. All participants first reached a reduced weight on semaglutide during a 20-week run-in; they were then randomized to continue or switch to placebo. The continuation group maintained and extended their loss; the placebo group regained substantially.
This is the single most important trial for the cost-and-duration conversation: it demonstrates, in a randomized design, that the weight effect depends on continued treatment. That is why GLP-1 therapy is usually a long-term commitment rather than a short course.
STEP 4's identifiers are fully verified (JAMA 2021; DOI and PMID linked below), and it pairs naturally with SURMOUNT-4 as the tirzepatide analogue.
| Trial | Population | N | Result | Citation | Identifiers |
|---|---|---|---|---|---|
| STEP 1 | Adults with overweight/obesity, without diabetes | 1,961 | Mean weight change about -14.9% vs -2.4% placebo | Wilding JPH, et al. N Engl J Med 2021;384:989-1002 | NCT03548935 · PMID 33567185 · DOI |
| STEP 4 | Adults with overweight/obesity after 20-week run-in | 803 | Continuing lost further weight; switching to placebo regained substantial weight | Rubino D, et al. JAMA 2021;325(14):1414-1425 | NCT03548987 · PMID 33755728 · DOI |
Identifiers verified against ClinicalTrials.gov and PubMed. Machine-readable dataset: studies.json · studies.csv.
What remains unknown
Across this topic, the recurring gap is the same: the randomized evidence describes approved products at studied doses, while compounded and deliberately sub-therapeutic 'microdose' use sits outside it. This page marks that boundary rather than blurring it.
What the evidence shows
- The cited trials carry verified NCT, PMID, and DOI identifiers and studied approved products at labeled doses.
- Findings here reflect those trials' actual results and populations.
What the evidence does not show
- That a compounded or microdose version reproduces these results — it has not been studied.
- That approved-product evidence transfers automatically to compounded formulations.
Frequently asked questions
What did STEP 4 show?
Continuing semaglutide maintained weight loss; switching to placebo led to substantial regain.
Why does STEP 4 matter?
It's the randomized basis for treating GLP-1 therapy as long-term.
Related: evidence database · research hub. Compounded medications are not FDA approved.
Not medical advice. This page is consumer education, not medical advice, and does not recommend a dose, product, or provider. Compounded medications are not FDA approved. Talk with a licensed clinician about your situation.