Both trials establish effects against placebo within their own study. Their headline percentages cannot, by themselves, isolate the difference between the medicines.
There is no randomized bridge between these two study populations.
Start with the study design
STEP 1 and SURMOUNT-1 are often placed together because they studied substantial weight reduction with weekly injectable treatments. The temptation is to compare the largest percentages and stop reading. A better starting point is to ask what participants were randomized to within each experiment.
STEP 1 assigned 1,961 adults without diabetes to semaglutide 2.4 mg or placebo for 68 weeks, with lifestyle intervention. SURMOUNT-1 assigned 2,539 adults without diabetes to tirzepatide 5, 10 or 15 mg, or placebo, for 72 weeks. Both used double-blind randomized designs and enrolled adults meeting specified BMI and weight-related condition criteria.[1] [2]
In each trial, randomization helps make the assigned treatment the main systematic difference between groups. It does not randomize people from one trial against people in the other. Placing the studies in adjacent columns does not create the experiment that was never conducted.
| Design detail | STEP 1 | SURMOUNT-1 |
|---|---|---|
| Study medicine | Semaglutide | Tirzepatide |
| Weekly target doses | 2.4 mg | 5, 10 or 15 mg |
| Duration | 68 weeks | 72 weeks |
| Randomized participants | 1,961 | 2,539 |
| Control | Placebo | Placebo |
| Funding | Novo Nordisk | Eli Lilly |
A percentage needs its context
The principal reported mean changes were −14.9% with semaglutide and −2.4% with placebo in STEP 1. In SURMOUNT-1 they were −15.0%, −19.5% and −20.9% across the ascending tirzepatide doses, versus −3.1% with placebo. These are study results at their respective endpoints, not a single dose-matched contest.[1] [2]
Duration matters because week 68 and week 72 are different observation windows. Dose matters because the three tirzepatide groups should not be compressed into one undifferentiated number. The control-group results matter because they remind us that the studies did not produce identical background changes.
Even subtracting each placebo result does not fully repair the comparison. That calculation may help describe each trial, but it cannot guarantee that all differences in populations, procedures and follow-up have been removed. A reader should resist replacing one oversimplified leaderboard with a slightly more elaborate one.
The SURMOUNT-5 head-to-head results answer the direct comparison more clearly. Our retatrutide versus tirzepatide analysis shows the limits of comparing separate trials.
Ask what happened to missing outcomes
A trial average is also a statistical answer to a defined question. How did investigators handle participants who stopped study medicine? Did the analysis aim to describe assignment to treatment regardless of discontinuation, or an effect under continued treatment? What assumptions filled gaps when a measurement was unavailable?
These questions can change how two percentages should be interpreted even when both appear under a heading such as “average weight loss.” An analysis restricted to people who remain on treatment may address a different decision from an analysis intended to include the consequences of stopping. Neither description should be silently substituted for the other.
For a consumer comparison, the honest response to an unspecified statistic is to ask for the paper and the named analysis. A dramatic percentage without the population, endpoint and treatment description is not enough information to evaluate the claim. More decimal places cannot restore missing context.
Use older trials for the right job
These trials remain important evidence that the studied treatments outperformed their own placebo groups. Their value is not erased by the fact that they were separate studies. They are useful for understanding study regimens, the range of measured responses and what researchers considered important enough to assess.
For a direct comparison between the studied semaglutide and tirzepatide regimens, SURMOUNT-5 is a more relevant design because it randomized the medicines within the same trial. It still has boundaries: it was open-label, used specified maximum tolerated doses, and studied adults without diabetes.[3]
The consumer decision is therefore a sequence. Match the evidence to the clinical question, discuss suitability, identify the exact product, and only then compare the service. A provider can accurately repeat a trial result while leaving the actual offer unclear. Ask whether the quote covers the formulation discussed, whether medication is separate, and how dose-related changes affect the bill.
Our medicine comparison guide separates the basic treatment differences. The provider comparison addresses the different question of care access and published costs. Read the dropout analysis before treating a trial's adverse-event percentage as your personal risk.