Why is the real-world difference so much smaller?

Mostly duration. Six months is roughly a third of the head-to-head trial's 72 weeks, and the difference between these two agents accumulates while doses escalate and continues afterwards. Dose attainment also differs - fewer tirzepatide patients reached the higher doses in the US cohort than semaglutide patients reached theirs.

Can an observational study establish which drug is better?

Not on its own. Propensity-score methods adjust for what was measured, and clinicians choose treatments for reasons that are often not in the record. Their value here is convergence: they agree in direction with a randomised trial and an indirect comparison, and designs with different weaknesses agreeing is stronger than any of them alone.

Does this apply to research material sold here?

No. Semaglutide and tirzepatide are licensed medicines prescribed and monitored by clinicians. Material supplied for laboratory research is not a version of either, is not an alternative to either, and the comparisons described here say nothing about it.

Which figure should be quoted?

Whichever one matches the question, with its duration attached. For what the drugs can achieve at maximum tolerated dose over more than a year, the randomised figure. For what a typical six months of routine prescribing produced in one US dataset, the cohort figure. Quoting either without its follow-up period is where the confusion starts.

Research & Regulatory News

The Gap Is Smaller Outside the Trial, and It Widens With Time

JMWritten & reviewed by Jack Muncaster · Founder, UK PeptidesLast reviewed 2026-08-235 cited sources

Two 2026 observational studies compared tirzepatide with semaglutide outside a trial setting. Both found tirzepatide ahead, and both found a smaller gap than the randomised head-to-head trial did, which is what the difference in follow-up duration and dose attainment would predict.

Key facts

US cohort: Truveta electronic health records, 2,396 on-treatment patients
US cohort split: 1,003 tirzepatide, 1,393 semaglutide
US follow-up: 6 months
US result: -11.15% vs -8.83%; adjusted difference -2.32 points (95% CI -3.17 to -1.48)
Dose attainment: 67.7% of semaglutide patients at 1.7 mg or above; 42.4% of tirzepatide patients at 10 mg or above
Chinese cohort: 183 adults, single centre, 3 months
Chinese result: -2.87 kg difference (95% CI -4.41 to -1.32)
Randomised head-to-head: SURMOUNT-5: -20.2% vs -13.7% at 72 weeks, 751 randomised

Two observational studies, one direction

SURMOUNT-5 established a randomised head-to-head result: tirzepatide -20.2% against semaglutide -13.7% at 72 weeks in 751 adults with obesity and without diabetes, a difference of 6.5 percentage points. Two 2026 observational studies asked whether the same ordering appears in practice. A US retrospective cohort using Truveta de-identified electronic health records followed 2,396 adults without diabetes who initiated either drug between December 2023 and June 2024 and remained on treatment for six months. A single-centre Chinese cohort followed 183 adults for three months. Both found tirzepatide ahead. Neither found anything like 6.5 points.

The US numbers, and the method behind them

Mean six-month weight reduction was -11.15% with tirzepatide against -8.83% with semaglutide, an adjusted difference of 2.32 percentage points (95% CI -3.17 to -1.48) from propensity-score weighted regression, with a modified intention-to-treat sensitivity analysis reaching consistent conclusions. Higher proportions of tirzepatide patients reached the 5%, 10%, 15% and 20% reduction thresholds, and reductions in BMI, blood pressure and HbA1c were also greater. Propensity-score weighting adjusts for measured differences between people who were prescribed one drug rather than the other; it cannot adjust for the unmeasured reasons a clinician chose one over the other, which is the standing limitation of every non-randomised comparison of treatments.

The detail that changes the interpretation

The US cohort reports something easy to skip: 67.7% of semaglutide patients reached 1.7 mg or above, while only 42.4% of tirzepatide patients reached 10 mg or above. In other words the comparison ran with semaglutide patients more likely to be on a higher dose than tirzepatide patients, and tirzepatide still came out ahead. That cuts against the obvious objection - that the real-world tirzepatide advantage might be an artefact of more aggressive titration - and points the other way: if tirzepatide patients had titrated as far, the observed gap would likely have been larger, not smaller.

Why the gap grows with duration

Line the three studies up by follow-up and a pattern appears. At three months the Chinese cohort found a 2.87 kg difference. At six months the US cohort found 2.32 percentage points. At 72 weeks the randomised trial found 6.5 percentage points. This is what a divergence that accumulates during and after titration looks like. Both drugs escalate slowly, so early in treatment the two arms are closer to each other than their maximum doses would suggest, and the separation widens as escalation completes and the effect plateaus at different levels. Quoting a six-month real-world figure as though it contradicts a 72-week trial figure is a category error; they measure different points on the same curve.

What the observational studies add that the trial cannot

SURMOUNT-5 was open-label, ran in a selected population that met eligibility criteria and consented to 72 weeks of study, and supported adherence in ways ordinary care does not. The observational cohorts describe people in routine practice with the discontinuation, dose interruption and supply problems that go with it. The US cohort restricted its primary analysis to patients who adhered, which removes some of that - and is why the modified intention-to-treat sensitivity analysis matters. Between them, the trial says what the drugs can do and the cohorts say what tends to happen, and the honest description of the field needs both.

The limits worth stating

The Chinese cohort is 183 people at a single centre over three months, comparing semaglutide 1.7 mg with tirzepatide 5 mg - fixed, non-maximal doses - and it reports its primary difference in kilograms rather than percent. The US cohort is larger and better adjusted but still observational, restricted to one country's electronic records over a specific enrolment window, and confined to people without diabetes. Neither can support a causal claim of the kind a randomised trial supports. Both are licensed medicines that this site does not supply, and nothing sold here is an alternative to either.

Quick reference

Study	Design	n	Follow-up	Difference favouring tirzepatide
Chinese single-centre cohort	Retrospective observational	183	3 months	-2.87 kg
US Truveta cohort	Retrospective, PS-weighted	2,396	6 months	-2.32 percentage points
SURMOUNT-5	Randomised, open-label	751	72 weeks	-6.5 percentage points
Network meta-analysis	Indirect, 6 trials	9,355	Varies	-6.26 percentage points

Extended research context

The Research & Regulatory News deep dive

Deep dive: why 2026 was the year the incretin field split in two

For a decade every meaningful GLP-1 medicine was a peptide, and every one of them was injected. 2026 broke that pattern in both directions at once. In August the MHRA authorised orforglipron, a small molecule with no peptide bonds that works as an ordinary daily tablet. Three months earlier, retatrutide's TRIUMPH-1 reported a 28.3% mean weight reduction — the largest figure yet from a single molecule, and achievable only with a peptide capable of engaging three receptors at once. The field is not converging on one answer; it is separating into a convenience track and a magnitude track, and those tracks have different chemistry.

Deep dive: what a marketing authorisation actually means

An authorisation is granted to a specific product, in a specific formulation, for a specific indication, by a specific regulator. It is not a statement about a compound class and it does not transfer. Orforglipron being licensed in the UK tells you nothing about the legal or regulatory status of any other incretin, and nothing at all about compounds that remain investigational. Authorisation is also separate from funding: a licensed medicine is not automatically available on the NHS, which requires a further NICE appraisal.

Deep dive: reading trial results without being misled

Headline percentages are the least transferable part of a trial. A figure is only meaningful alongside its population, its duration, its comparator and its dropout rate. TRIUMPH-1's 28.3% came from an 80-week study in a relatively uncomplicated obesity population; TRIUMPH-3's 22.6% came from adults with severe obesity and established cardiovascular disease, and the gap between those two numbers is mostly population, not potency. Discontinuation rates deserve the same attention as efficacy: 11.3% of the TRIUMPH-1 12 mg arm stopped due to adverse events against 4.9% on placebo, and that figure is routinely dropped from summaries.

Research applications

▸Tracking regulatory status of investigational incretin compounds
▸Understanding the difference between authorisation, NICE appraisal and NHS availability
▸Comparing peptide and non-peptide receptor agonist pharmacology
▸Interpreting Phase 3 topline releases before peer-reviewed publication
▸Verifying trial identity against ClinicalTrials.gov registrations

Handling checklist

✓Check the compound named in a trial registration matches the compound being discussed
✓Confirm the NCT identifier resolves to the acronym being cited
✓Read topline press releases as preliminary until peer-reviewed publication
✓Separate the trial population from the headline percentage before comparing studies
✓Treat authorisation in one jurisdiction as saying nothing about status in another

Common research-handling mistakes

Learnt from thousands of researcher orders across our UK labs.

✗ Treating a positive Phase 3 as approval

Fix: Filing begins a review that commonly takes a year or more and can end in a request for more data.

✗ Comparing weight-reduction percentages across different trials

Fix: Population, duration and comparator differ; the numbers are not interchangeable.

✗ Assuming one incretin's approval legitimises another compound

Fix: Authorisations are product-specific and do not transfer between compounds.

✗ Citing a TRIUMPH number without checking the NCT identifier

Fix: Verify against ClinicalTrials.gov — the numbering has been widely misreported.

✗ Reading research material as an alternative to a licensed medicine

Fix: Research material is supplied for laboratory use only and is not a substitute for anything prescribed.

Continue researching

Peer-reviewed guides, comparators and matched reference materials.

Frequently asked questions

Why is the real-world difference so much smaller?: Mostly duration. Six months is roughly a third of the head-to-head trial's 72 weeks, and the difference between these two agents accumulates while doses escalate and continues afterwards. Dose attainment also differs - fewer tirzepatide patients reached the higher doses in the US cohort than semaglutide patients reached theirs.
Can an observational study establish which drug is better?: Not on its own. Propensity-score methods adjust for what was measured, and clinicians choose treatments for reasons that are often not in the record. Their value here is convergence: they agree in direction with a randomised trial and an indirect comparison, and designs with different weaknesses agreeing is stronger than any of them alone.
Does this apply to research material sold here?: No. Semaglutide and tirzepatide are licensed medicines prescribed and monitored by clinicians. Material supplied for laboratory research is not a version of either, is not an alternative to either, and the comparisons described here say nothing about it.
Which figure should be quoted?: Whichever one matches the question, with its duration attached. For what the drugs can achieve at maximum tolerated dose over more than a year, the randomised figure. For what a typical six months of routine prescribing produced in one US dataset, the cohort figure. Quoting either without its follow-up period is where the confusion starts.

Primary sources & clinical trials

Peer-reviewed research and registered trials from PubMed, ClinicalTrials.gov, PubChem, FDA and NIH. All links open in a new tab and point to the primary source, so every claim can be verified at origin.

Written and reviewed by

Jack Muncaster · Founder, UK Peptides

Jack founded UK Peptides in Manchester after repeatedly receiving research compounds with missing or recycled paperwork. He is responsible for supplier selection, batch release decisions and the content published in this research library. Every article here is sourced to primary literature and every product page to a signed third-party certificate.

How we write and check this content Report a factual error in this article

More Research & Regulatory News articles

Popular across the research hub

One flagship guide from every other research category — keep exploring.

← Back to Forum