STEP 5 trial results showed average weight loss of 15.2% with semaglutide versus 2.6% with placebo, an inactive injection, over two years. Both groups received support with eating and physical activity. The average weight curve flattened near week 60, then stayed broadly steady through week 104.
That matters if the worry is whether the early loss disappears while treatment continues. In this trial, the group average held. It doesn't mean everyone lost the same amount, or that two years settles lifelong use.
What question did the two-year semaglutide study ask?
Could weight loss with continued semaglutide treatment last for 104 weeks? W. Timothy Garvey and colleagues reported STEP 5 in Nature Medicine in 2022.
Earlier studies were shorter. The STEP 1 results guide covers the 68-week comparison.
Who enrolled, and what treatment did they receive?
STEP 5 assigned 304 adults by chance: 152 to semaglutide and 152 to placebo. Neither participants nor investigators knew the assignment. The trial registration, NCT03693430, records the planned study.
Participants were at least 18; their average age was 47.3. They needed a BMI, a measure of weight relative to height, of at least 30. A BMI of at least 27 qualified with a weight-related condition, such as high blood pressure.
People with diabetes were excluded, alongside people with recent weight-loss medication use or previous or planned weight-loss surgery. Everyone reported an unsuccessful dietary effort to lose weight.
Women made up 77.6% of the group, and 93.1% were White. These results therefore represent some people much better than others.
The trial used weekly semaglutide injections, increasing over 16 weeks toward 2.4 mg. Both groups received counseling every four weeks, a reduced-calorie eating plan and a physical-activity goal. These were study conditions; an individual's prescriber sets the actual dose.
What did researchers measure?
The co-primary endpoints, the two main measurements chosen beforehand, were percentage weight change and the share losing at least 5% at week 104.
An estimand defines the treatment effect an analysis aims to measure. The estimand explainer explains why one trial can have two valid headline numbers.
The treatment policy estimand included people after treatment stops or additional weight-management treatment. The trial product estimand estimated what would happen if everyone continued assigned treatment without additional weight-management treatment. Both used statistical estimates for missing measurements; the second isn't simply an average of finishers.
STEP 5 trial results: both calculations
The main analysis found 15.2% average loss with semaglutide versus 2.6% with placebo. The gap was 12.6 percentage points, meaning the difference between the two percentages.
| Analysis at week 104 | Semaglutide weight change | Placebo weight change |
|---|---|---|
| Treatment policy: includes treatment stops | −15.2% | −2.6% |
| Trial product: assumes continued treatment | −16.7% | −0.6% |
The larger semaglutide number comes with a condition: continued treatment. Neither figure is a personal target.
About half of semaglutide participants with a week-104 measurement lost at least 15%. About one in three lost at least 20%.
| Loss from starting weight | Semaglutide | Placebo |
|---|---|---|
| At least 5% | 77.1% | 34.4% |
| At least 10% | 61.8% | 13.3% |
| At least 15% | 52.1% | 7.0% |
| At least 20% | 36.1% | 2.3% |
These are observed proportions among 144 semaglutide and 128 placebo participants with measurements, regardless of treatment stops. The 20% threshold was a supporting outcome outside the main sequence of statistical tests.
The curve plateaued around week 60. Average semaglutide weight loss was 15.6% at week 52 and 15.2% at week 104. Holding that average differs from continuing to lose weight indefinitely.
What side effects and treatment stops occurred?
About four in five semaglutide participants reported stomach or bowel problems at some point, versus roughly half on placebo. Most were temporary and mild to moderate.
These selected rows reproduce participant percentages from the paper's adverse-event table. An adverse event is a health problem recorded during a study, whether or not treatment caused it.
| Recorded event or treatment stop | Semaglutide | Placebo |
|---|---|---|
| Any stomach or bowel disorder | 82.2% | 53.9% |
| Nausea | 53.3% | 21.7% |
| Diarrhea | 34.9% | 23.7% |
| Constipation | 30.9% | 11.2% |
| Vomiting | 30.3% | 4.6% |
| Stopped treatment because of adverse events | 5.9% | 4.6% |
| Stopped because of stomach or bowel disorders | 3.9% | 0.7% |
| Serious adverse events | 7.9% | 11.8% |
Serious events include problems requiring hospitalization. Having symptoms and stopping treatment are different outcomes; these percentages also don't mean symptoms lasted continuously for two years.
Overall, 92.8% completed the trial's final safety visit, while 79.9% remained on assigned treatment at the end-of-treatment visit. Completing follow-up isn't the same as continuing medication.
What later evidence adds, and what remains open
The authors highlighted the small proportion of non-White participants and the mostly female sample. Study counseling may also differ from ordinary care. Novo Nordisk funded STEP 5 and designed, oversaw and analyzed the trial.
Two years can show maintenance during that period. It cannot establish decades of safety, identify every rare harm or determine how long one person needs treatment. STEP 5 also didn't test a planned withdrawal strategy; the STEP 4 results guide covers switching to placebo after initial treatment.
A later SELECT analysis in Nature Medicine, 2024, followed weight through four years. Average loss was 10.2% with semaglutide versus 1.5% with placebo at week 208. Its participants had established cardiovascular disease, meaning heart or blood-vessel disease, so it wasn't simply a longer STEP 5.
As of October 1, 2026, the day this article was written, a comparison published September 28 in the International Journal of Obesity had compared semaglutide with tirzepatide using both analysis approaches. It found broadly similar conclusions across approaches, with some differences in which results cleared the statistical cutoff for distinguishing a treatment difference from chance. Lilly funded it. It compared separate STEP 1 and SURMOUNT-1 trials, rather than assigning people directly between the medicines, and didn't extend STEP 5's follow-up.
Tirzepatide is a separate medication a provider may consider. This comparison cannot decide which treatment fits an individual. The broader semaglutide weight-loss results guide covers the other populations and comparisons.
What this means for a compounded prescription
STEP 5 tested the study semaglutide product, not Promise's preparation. Through Promise, semaglutide is dispensed as a compounded medication, prepared by a licensed pharmacy for an individual prescription, which is different from an FDA-approved product: the formulation offered here is not FDA-approved.
The FDA's compounding explanation, marked “Content current as of: 06/22/2026,” says the agency doesn't review these preparations' safety, effectiveness or quality before marketing. STEP 5's averages cannot establish the results of that specific preparation.
A licensed provider may still prescribe a compounded formulation; that decision is between the patient and the doctor. At Promise, a licensed provider reviews every request, and not everyone qualifies. Medical history, response and side effects belong in the conversation about continuing treatment.
