Tirzepatide: The Trials Are Real, Which Is Why Somebody Copied One

Every other entry here has the same shape. Interesting mechanism, a pile of rodent work, almost nothing in humans.

Tirzepatide breaks it completely. Two FDA approvals, and a numbered series of phase 3 trials, placebo controlled and active comparator controlled, in the New England Journal of Medicine, the Lancet and JAMA.

So this entry has a different job. The evidence is not the weak point. What has attached itself to the evidence is.

The record that carries Lilly's protocol code

On 31 July 2026, Morgan McSweeney published an investigation into a cluster of eight ClinicalTrials.gov registrations sharing one sponsor, Hudson Biotech. All eight read as recruiting, all start in February 2026, all name a single site at Peking University Shenzhen Hospital. (Dr Noc)

I ran the sponsor query and got the same eight. Six are grey market research peptides. Two reuse Eli Lilly compound codes. Three say in their own text that they are examples, which the BPC-157 entry here covers. The tirzepatide record is the worst of them.

NCT07481747 is a clone of SURMOUNT-1. Submitted 8 March 2026, recruiting, enrolment 2,539, one site, a contact at a beijing-biotech.com address. Its official title is character for character identical to the genuine record. Its enrolment figure is what the real trial completed with, recycled as a forward looking estimate. Its design is the real design.

Then the field that settles it. The org study ID reads I8F-MC-GPHK(b). I8F-MC-GPHK is Eli Lilly's own internal protocol code for SURMOUNT-1, sitting as a secondary identifier on Lilly's genuine record. The clone carries it with "(b)" appended. Query the registry for that string and exactly two records come back. Lilly's, and this one.

A counterfeit note can be very good and still fail on one thing, which is that the serial number printed on it belongs to a note somebody else is already holding. Copying the protocol code made this record more convincing on its face. It also produced the one search that returns two answers where there should be one.

And nothing in it warns you. I read the brief summary and the detailed description in full. Neither contains the words fictional, example, mock, simulated or synthetic. Both read as ordinary protocol prose. Scanning an entry for a disclaimer is not a check. The only screen that catches this one is the sponsor.

Do not cite NCT07481747.

What it is

A 39 amino acid synthetic peptide that activates two receptors at once, the GIP receptor and the GLP-1 receptor. Built on the GIP sequence, carrying a C20 fatty di-acid, the piece that stretches it out to once weekly subcutaneous dosing. (PMID 30473097)

The dose range everything downstream rests on came from a phase 2 trial in type 2 diabetes: 318 patients, 26 weeks, tirzepatide at 1, 5, 10 and 15 mg weekly against dulaglutide 1.5 mg and placebo. (PMID 30293770) Hold on to that 1 mg arm.

The diabetes trials

SURPASS is five lettered phase 3 trials plus an outcome trial: monotherapy against placebo (PMID 34186022), against insulin degludec (PMID 34370970), against insulin glargine at raised cardiovascular risk (PMID 34672967), and against placebo on top of titrated glargine (PMID 35133415).

SURPASS-2 is the one to read closely. 40 weeks, open label, 1,879 adults on metformin, against semaglutide 1 mg. HbA1c fell 2.01, 2.24 and 2.30 percentage points against 1.86, non-inferior and superior at every dose, with additional weight loss of 1.9, 3.6 and 5.5 kg. (PMID 34170647)

SURPASS-CVOT put 6,586 people on tirzepatide against 6,579 on dulaglutide and found it non-inferior for the composite of cardiovascular death, myocardial infarction or stroke. (PMID 41406444) Summaries upgrade that. Active comparator, not placebo, and superiority was not met. SURPASS-PEDS took it into ages 10 to under 18, maximum dose 10 mg weekly rather than the adult 15 mg. (PMID 40975112)

The obesity trials

SURMOUNT-1 is the trial the clone copied. 72 weeks, double blind, placebo controlled, adults with obesity or overweight without type 2 diabetes. Weight change at week 72 was 15.0% at 5 mg, 19.5% at 10 mg and 20.9% at 15 mg, against 3.1% on placebo. A 20% or greater reduction was reached by 57% at 15 mg against 3% on placebo. (PMID 35658024)

SURMOUNT-2, the same design in people who also had type 2 diabetes, reports 12.8% and 14.7% against 3.2%. (PMID 37385275) You will more often see 13.4% and 15.7%, from Lilly's press release. Both are real. Those are the efficacy estimand and these the treatment regimen estimand, so say which you are quoting.

SURMOUNT-3 began only after a 12 week intensive lifestyle lead-in, measuring what the drug adds on top of a real behavioural programme. (PMID 37840095) SURMOUNT-4 is the one people skip: after a 36 week lead-in, participants were randomised to continue or switch to placebo, and withdrawal produced substantial weight regain. (PMID 38078870)

SURMOUNT-5 ran head to head for 72 weeks at maximum tolerated doses of each drug. 20.2% against 13.7% for semaglutide. (PMID 40353578) The cleanest comparison anyone has run, and it belongs beside the semaglutide entry here, where sourcing is the whole story.

Two trials measured something other than weight. SURMOUNT-OSA cut the apnoea hypopnoea index by 25.3 events per hour against 5.3 in people not using PAP, and by 29.3 against 5.5 in those already on it. (PMID 38912654) SUMMIT, in heart failure with preserved ejection fraction and obesity, found lower risk of cardiovascular death or worsening heart failure across 731 patients. (PMID 39555826)

What the label says

Zepbound is indicated, alongside a reduced calorie diet and increased physical activity, for long term weight reduction in adults with obesity or overweight plus a weight related comorbid condition, and for moderate to severe obstructive sleep apnoea in adults with obesity. Mounjaro is indicated as an adjunct to diet and exercise for glycaemic control in adults and paediatric patients 10 years and older with type 2 diabetes, capped at 15 mg weekly in adults but 10 mg in paediatrics. Both labels were revised 4/2026.

Both carry a boxed warning for thyroid C-cell tumours, explicitly species specific. The tumours were seen in rats, human relevance is undetermined, and both are contraindicated with a personal or family history of medullary thyroid carcinoma or MEN 2.

Label dosing, reported as label rather than as advice: 2.5 mg subcutaneously once weekly for four weeks, then 2.5 mg increments.

Where the microdosing convention runs out

The convention that circulates defines a microdose as any weekly dose below the 2.5 mg label starter, typically 0.25 to 1.5 mg. I found those numbers only on vendor and telehealth pages, which are marketing rather than evidence.

One correction first, because the usual version of this argument overreaches. Randomised, placebo controlled Phase 1 data below 1 mg do exist. Lilly's first-in-human study (Coskun 2018, NCT02759107) gave single doses from 0.25 mg and four weeks of weekly dosing from 0.5 mg, but those parts measured pharmacokinetics and short-term safety, not weight loss. No randomised trial has evaluated a sub-1 mg weekly dose as a weight-management regimen.

Here is the gap. The 1 mg arm of the phase 2 trial is the lowest tirzepatide dose ever carried through a randomised trial with a sustained dosing weight endpoint, and it produced only 0.9 kg of weight loss over 26 weeks in people with type 2 diabetes. Randomised exposure at 0.25 to 0.5 mg exists only in Lilly's first-in-human Phase 1 study, which ran for a single dose or four weeks and was not designed to measure weight loss. Community protocols promising meaningful weight loss at 0.5 to 1.5 mg weekly are extrapolating past the only randomised weight data anywhere near their dose range, and that data is unflattering. (PMID 30293770)

Look closer. HbA1c fell 1.06 percentage points at 1 mg against 1.94 at 15 mg, so on blood sugar the bottom dose held more than half the effect of the top. On the scale it gave 0.9 kg, at the bottom of a range reaching 11.3 kg. The two effects do not shrink together.

An airliner throttled back still runs its lights, its radios and its cabin air perfectly well. Every system you can check reports normal. None of it tells you the thing can leave the ground.

Nor did side effects vanish at the bottom. Gastrointestinal events were dose related at 23.1% for 1 mg, 32.7% for 5 mg, 51.0% for 10 mg and 66.0% for 15 mg, against 9.8% on placebo. The practice has now been criticised in the peer reviewed nursing literature, which frames it as patient anecdote driven and flags pen manipulation, compounded vials and medication sharing. (PMID 42201545)

The compounding fight

One note before the dates. FDA's own copies of these documents returned errors to every fetch I attempted, and the court opinion below was initially under seal. This sequence is therefore reported from industry summaries rather than from the primary record, the Federal Register items excepted.

FDA first removed tirzepatide from the shortage list in early October 2024. The Outsourcing Facilities Association sued on 7 October 2024 (announced by press release the following day) in the Northern District of Texas, and on 11 October 2024, on FDA's own motion, the court remanded the decision to the agency.

On 19 December 2024 FDA issued a declaratory order reaffirming the shortage resolved across all six strengths. 503A pharmacies had 60 days to wind down, to 18 February 2025. 503B outsourcing facilities got 90 days, to 19 March 2025, because they are FDA registered and subject to cGMP. On 5 March 2025 Judge Mark Pittman denied the association's motion for a preliminary injunction, and legal compounding ended.

Then FDA moved to close the pathway permanently. A Federal Register notice published 1 May 2026 under docket FDA-2018-N-3240 proposed not to include semaglutide, tirzepatide and liraglutide on the 503B Bulks List, having found no clinical need. It rejects every argument the compounders made, including that hyper-responders need sub-label doses, noting that approved products already offer lower strengths. Patient preference, it states plainly, is not clinical need.

Separately, the US International Trade Commission issued a general exclusion order on 9 April 2025 in Investigation No. 337-TA-1377, brought by Eli Lilly. It bars importation of products containing tirzepatide and of "products purporting to contain tirzepatide". That last phrase is a federal agency recording, inside an enforcement order, that some imported product may not contain tirzepatide at all.

What is actually in compounded product

The strongest published analytical finding needs the loudest disclosure attached. A 2026 paper reports a previously unidentified impurity formed by reaction between tirzepatide and certain vitamin B12 analogues, across ten samples from compounding pharmacies, medspas and telehealth networks. (PMID 42010938) All five authors are Eli Lilly employees. That is manufacturer run testing of a competitor product category. The paper is honest about its ceiling, saying the "clinical effects of this impurity are unknown."

Three other datasets sit around it. FAERS, 2018 to 2024: 707 of 81,078 GLP-1 receptor agonist reports involved compounded products, with far higher reporting odds ratios for preparation errors at 48.92, contamination at 19.00 and hospitalisation at 2.35. That is disproportionality analysis pooled across three drugs, so it cannot isolate tirzepatide or establish causation. (PMID 40285721)

A survey of 33 compounded products advertised in early 2025, one third of them tirzepatide, found single ingredient products were 82% sublingual and 18% orally disintegrating tablets, neither an approved route, with only 18% reporting beyond use dating. (PMID 41689811)

The market did not disappear afterwards. Among 75 weight loss clinics and medspas in West Virginia and Oklahoma trading August to October 2025, 56% offered GLP-1 products combined with B vitamins. Of 23 compounding suppliers, 4 of the 21 whose licensure could be determined (19.0%) were not licensed for sterile compounding, 3 of 22 (13.6%) had faced state level disciplinary action, and one had multiple FDA warning letters. (PMID 42467450)

One outcome dataset is dominated by compounded tirzepatide and gets cited as though it settles the question. 1,166 telehealth patients completing at least 52 weeks, 91.7% on compounded product, mean weight change 8.80% at 12 weeks and 26.54% at 72. It is retrospective, unblinded, has no control arm, does not analyse compounded and branded separately, and was produced by the programme it evaluates. It is not evidence that compounded performs like branded. (PMID 41438798)

What nobody knows

What a sub-1 mg weekly dose does to weight over time. No randomised trial has evaluated one as a weight management regimen.

What the B12 reaction impurity does to people. The paper that found it says so itself, and every author works for the branded manufacturer.

My take

Tirzepatide is the strangest entry in this reference, because for once the science is not the problem.

Somebody ran the trials. Big, long, against active comparators, and they published a randomised withdrawal trial showing the weight comes back, which is not a result a sponsor volunteers unless the programme is fundamentally honest. You can argue about magnitude. You cannot argue that nobody looked.

What has grown around that evidence moves in two directions. Downward, into doses tested exactly once, that performed poorly when they were. Sideways, into compounded and imported product no trial ever used, where the regulator's own language has shifted to "purporting to contain."

Then there is the registry record. Every other fake I have looked at in this project gave itself away in its own text. This one does not. It was built by copying something genuine whole, file number included, which made it stronger everywhere except the single place a duplicate becomes visible. The evidence being real is what made it worth stealing.

Have you seen a clinic or telehealth site quote SURMOUNT-1's 20.9% while selling something that is not branded tirzepatide? I would like to know which one, because that specific move is why this entry exists.

---

This content is for educational and informational purposes only and is not medical advice. Peptides discussed are research compounds and may not be approved for human use. Nothing here should be used to diagnose, treat, cure, or prevent any disease. Always consult a qualified healthcare professional before starting any peptide, supplement, or protocol. Individual responses vary. Do not self-administer compounds without proper medical supervision.