Tirzepatide vs semaglutide: what the only head-to-head trial found

One trial compared them directly. It was open-label, and it was funded by the company that makes the winner.

The analysis

Design. Phase 3b, open-label, randomized head-to-head. 751 adults with obesity (BMI ≥30, or ≥27 with a weight-related comorbidity) and without type 2 diabetes, randomized 1:1 to maximum tolerated tirzepatide (10 or 15 mg) versus maximum tolerated semaglutide (1.7 or 2.4 mg), once weekly for 72 weeks.

Result. Least-squares mean body-weight change at week 72: −20.2% with tirzepatide (95% CI −21.4 to −19.1) versus −13.7% with semaglutide (95% CI −14.9 to −12.6), p<0.001 — about 47% greater relative weight loss, or 22.8 kg versus 15.0 kg. Tirzepatide was superior on the primary endpoint and all five key secondary endpoints. 31.6% of tirzepatide patients lost at least 25% of body weight, versus 16.1% on semaglutide.

Source: Aronne LJ et al., New England Journal of Medicine, May 11, 2025. NCT05822830.

Absolute versus relative: reading the number correctly

Trial results are usually reported as relative figures, because relative figures are larger and therefore more persuasive. A "20% reduction in cardiovascular events" sounds transformative. The absolute reduction in SELECT was from 8.0% to 6.5% — about 1.5 percentage points over roughly three years. Both statements describe the same result honestly; only one of them tells you what to expect for yourself.

The same applies to weight-loss figures. A mean reduction of 20.9% is a mean. Individual results in these trials ranged from substantial loss to none at all, and a mean tells you nothing about where you personally would land. Anyone quoting a trial average as a promise is misusing it.

Funding and conflicts of interest

Every pivotal trial in this field was funded by the company that manufactures the drug it tested. That is normal in pharmaceutical research and it does not make the results false — these are large, well-conducted, peer-reviewed studies. It does mean the funding belongs in the citation every time, particularly for head-to-head trials where the funder makes the winning drug. SURMOUNT-5 was funded by Eli Lilly and found Lilly's drug superior. The result is plausible and consistent with the separate trial programmes; the disclosure still belongs beside it.

Where this sits against the other evidence

No single trial should be read alone. The strength of the GLP-1 evidence base is that multiple independent trial programmes — SURMOUNT for tirzepatide, STEP for semaglutide, SCALE for liraglutide, SELECT for cardiovascular outcomes — point in a consistent direction across tens of thousands of participants. That consistency is what makes the class credible.

What that consistency does not do is extend to products the trials never tested. Every one of those programmes studied an FDA-approved subcutaneous injection. None studied a compounded preparation, a microdose regimen, or an orally disintegrating tablet. The evidence is strong exactly where it was collected and silent everywhere else, and the gap between those two things is where most of the marketing in this industry operates.

What to do about it

Three practical steps follow from everything above.

1. Check your insurance first. A covered brand prescription with a manufacturer savings card can cost roughly $25 a month, which beats every cash option discussed here.
2. Then price the manufacturer directly. LillyDirect and NovoCare sell brand GLP-1s for $149-$449. Several telehealth platforms resell the identical drugs at four to eleven times that.
3. Then, and only then, compare compounded programmes — on their ongoing total cost, medication plus any mandatory membership, at the dose you expect to maintain.

Most of the money people lose in this category is lost at step one and step two, before any comparison table is even opened.

Limitations of this analysis

Every page on this site should tell you where it stops being reliable. This one stops here.

Prices decay quickly. This is the fastest-moving data we publish. Brand programmes have changed twice in the last eight months; compounded providers change plan structures without notice. Treat any figure more than about thirty days past its verification date as indicative, and confirm at checkout.

Competitor pricing is reported, not captured by us. We hold dated captures for brand pricing and for NexLife. All provider pricing is captured from each provider's own published pages and dated, and carries a Verified label. Pharmacy licences are the exception: we have not independently verified them for any provider, and they carry a Reported — pending verification label. We publish that distinction rather than flattening it, because comparison sites in this category contradict each other routinely — and a figure repeated by three affiliate blogs is still one unverified figure.

We have not audited pharmacy licences. Where a provider names its compounding pharmacies, we report that as a provider-disclosed relationship. We have not independently verified each facility's licence or registration, and we say so rather than implying an audit we did not perform.

Advertised availability is not your availability. Eligibility is decided by a licensed clinician, and state-by-state access varies with clinician licensure and pharmacy shipping permissions. No page can promise you a price you will actually be offered.

We are commercially funded. The publisher and certain principals have financial relationships with some of the providers listed here, and we may earn a commission from provider links. That is disclosed in the footer of every page. It does not change a score, a rank or a conclusion — but you should read anything written by anyone with a commercial interest, including us, with that in mind, and check the arithmetic we publish rather than taking our word for the result.

Update history