Two approved incretin medicines, one direct comparison trial and a great deal of confusion in between. A source-checked reading of STEP 1, SURMOUNT-1, SURPASS-2 and the open-label SURMOUNT-5 comparison.
Semaglutide and tirzepatide are the two incretin-based medicines that changed the metabolic field, and they are constantly compared in ways the trial record does not support. Both are approved prescription medicines with regulator-reviewed labels — which already separates them from the unapproved research peptides discussed elsewhere on this site. What is far less clear in public discussion is what has actually been measured, in whom, and under what trial design.
Semaglutide is a GLP-1 receptor agonist: it acts on a single incretin receptor pathway. Tirzepatide acts on two — the GIP receptor and the GLP-1 receptor. That is a pharmacological difference, not automatically a clinical one; a second receptor target only matters to the extent trials show it does.
The two landmark placebo-controlled trials are frequently placed side by side as though they were a comparison:
STEP 1 (NEJM, 2021) studied once-weekly semaglutide in adults with overweight or obesity without diabetes, against placebo, alongside lifestyle intervention.
SURMOUNT-1 (NEJM, 2022) studied once-weekly tirzepatide in a comparable population, also against placebo.
Reading a percentage from one and a percentage from the other is not a comparison. The trials differ in enrolment period, populations, background support, treatment duration and estimand — the statistical definition of what "weight change" means when participants stop treatment. Cross-trial arithmetic of this kind is the most common error in coverage of this field.
Only a randomised trial that gives both drugs to comparable participants under the same protocol can support a comparative statement.
Two randomised head-to-head trials have been published:
SURPASS-2 (NEJM, 2021), in adults with type 2 diabetes, compared tirzepatide with semaglutide 1 mg on glycaemic outcomes, with body weight as a secondary endpoint. It is a diabetes trial, and its semaglutide dose is the diabetes dose — not the higher dose later used for obesity.
SURMOUNT-5 (NEJM, 2025; ClinicalTrials.gov NCT05822830) is the phase 3b trial that compared tirzepatide with semaglutide in adults with obesity and without type 2 diabetes. It was an open-label , controlled trial: participants and investigators knew which drug was given.
Open-label design is a genuine limitation and should be stated whenever the trial is cited. It does not invalidate the result, but expectation effects cannot be excluded in outcomes such as appetite, eating behaviour and tolerability reporting.
Across all four trials, the dominant adverse events in both drug classes were gastrointestinal — nausea, diarrhoea, vomiting, constipation — mostly reported as mild to moderate and concentrated during dose escalation. Discontinuation due to adverse events occurred in a minority of participants. These are trial-population figures under supervised escalation; they are not a personal risk estimate, and the approved product labels remain the authoritative safety documents.
It does not make either drug "better" in a general sense. Trial outcomes are averages in selected populations over fixed periods, measured on specific endpoints.
It does not transfer between populations. A diabetes trial result does not describe people without diabetes, and vice versa.
It says nothing about compounded or research-grade material. The trials studied manufactured, regulator-inspected medicines. The FDA has warned specifically about unapproved GLP-1 products sold for weight loss, where identity, purity and dose are unverified.
The honest summary is narrower than the discourse: both are approved, both work substantially better than placebo in their trial populations, one direct obesity comparison exists and it was open-label. Everything beyond that — individual response, long-term outcomes, comparative durability after stopping — is either still being studied or is a prescribing decision for a clinician with the label in front of them. This publication does not provide dosing, escalation schedules or sourcing information for either medicine.
The listed trial reports and the FDA communication are the primary material for this review. Each is linked to its DOI or official record so readers can check the population, design and endpoint definitions themselves.