Retatrutide: The Registry Record That Copies a Trial That Already Finished

01What it is

An agonist is a molecule that switches a receptor on. Retatrutide, developed as LY3437943, is one peptide with agonist activity at three receptors: those for glucose dependent insulinotropic polypeptide, glucagon like peptide 1, and glucagon. (PMID 35985340) Triple agonist is not the same as balanced agonist though. On Lilly's own numbers it is 8.9 fold greater than human GIP at the GIP receptor, 2.5 fold less than GLP-1 at its receptor, and 2.9 fold less than glucagon at its own.

Three notes do not make a chord you hear equally. The loud note is GIP.

02What the trials found

The pivotal obesity trial, the one the fake record copied, randomised 338 adults to once weekly retatrutide at 1, 4, 8 or 12 mg or placebo for 48 weeks. (PMID 37366315)

Look at the two middle rows at 24 weeks. Going from 8 mg to 12 mg bought 0.2 percentage points. Those arms separate only in the back half.

Blood sugar next. HbA1c is a blood test reflecting average blood sugar over about three months. Phase 2 also ran in type 2 diabetes, where HbA1c fell up to 2.02%. (PMID 37385280)

TRIUMPH-1, NCT05929066, is a real Lilly master protocol. 2,335 participants, testing 4, 9 and 12 mg. An estimand is the counting rule, and the efficacy one asks what happens in people who stay on the drug. At 80 weeks the efficacy estimand gave 19.0%, 25.9% and 28.3% weight loss against 2.2% on placebo, and the more conservative treatment regimen estimand 17.6%, 23.7% and 25.0%. Announced by the company on 21 May 2026 and not published in a peer reviewed journal.

03The record that copies a finished trial

ClinicalTrials.gov carries a retatrutide entry, NCT07467447. Sponsor Hudson Biotech. Phase 2. Status recruiting. One site, Peking University Shenzhen Hospital. Its official title is character for character identical to Eli Lilly's NCT04881760, and the description copies the architecture too: maintenance doses of 1, 4, 8 and 12 mg, 48 weeks of treatment, a Week 24 primary endpoint of percent change in body weight, the same alternative escalation schedules for the 4 mg and 8 mg arms. The real trial completed on 22 November 2022.

The credit for finding this cluster belongs elsewhere. On 31 July 2026, Morgan McSweeney published an investigation into eight registrations under that sponsor, all at the same single site, all submitted within three weeks of each other, all marked recruiting. (Dr Noc) Three say plainly what they are, the TB-500 entry opening "This fictional study is an example of a ClinicalTrials.gov-style record." Two others clone real Lilly trials. What this entry adds is one name: that piece noted the compound code LY3437943 among those two without saying which compound LY3437943 is. It is retatrutide.

I read the complete brief summary and detailed description of NCT07467447. Neither contains the word fictional, example, mock, simulated, hypothetical or illustrative. Not once. Any screen looking for the giveaway words passes this record through.

NCT07467447 is the one retatrutide registry entry that must never be cited. The other sixteen I checked are all Lilly registrations bar one investigator study at Massachusetts General.

04How it compares

Retatrutide gets sold beside the two drugs nearest to it, and their approvals get waved at it.

Lilly's own medical information is unqualified: retatrutide is investigational, approved in no country, and cannot be legally sold or marketed for human use. Lilly says it plans to file in the first quarter of 2027. There is exactly one legal route to it outside a trial, and it is not a shop. NCT07629401 is a Lilly expanded access record, physician initiated and case by case. Compounding is no way around it either, since retatrutide has no monograph and sits on no bulks list, though FDA's bulks list page would not load for me and that comes from FDA's letter to the Federation of State Medical Boards.

05Where the circulating doses came from

Convention, because that is what it is. No trial supplied drug the way vendor charts teach.

1

2

3

4

0.1

1

10

mg per week · log scale

derived from researchcommunity convention

1. 1The circulating starting range · 0.25 to 1.5 mg per weekCharts start at 0.25 to 0.5 mg weekly, some splitting the week across three injections. No published trial used sub 1 mg as a maintenance target, or split dosing. The phase 2 diabetes 0.5 mg arm is the only published sub 1 mg exposure.
2. 2The circulating maintenance target · 4 mg per weekThe steady state those charts aim at, and the lowest maintenance dose in phase 3.
3. 3The highest dose in any published trial · 12 mg per weekPhase 2 used 1, 4, 8 and 12 mg as maintenance, phase 3 used 4, 9 and 12 mg. The registry shows the 12 mg arm started at 2 mg, so the 1 then 4 then 8 then 12 ladder people copy is not what the trial did.
4. 4The dose in the one published case report · 20 mg per weekOne patient, not a norm. Both are vial numbers rather than trial doses: 10 mg is a standard vial, 20 mg is two.

06What could go wrong

The tolerability ceiling shows in phase 3. At 12 mg against placebo, nausea ran 42.4% versus 14.8% and vomiting 25.3% versus 4.8%. Discontinuation for adverse events climbed 4.1%, 6.9%, 11.3% across the three doses, against 4.9% on placebo. More than one in nine quit the top dose over side effects.

The only vial level numbers I found are vendor commissioned: a shop provided BulkGLP 10 mg vial came back 99.76% pure but measured 12.05 mg, 121% of the label, with no sterility screen. (Peptigrity) Anyone dosing off that label is off by a fifth.

The only peer reviewed report of harm I could find is one case. A 32 year old man who bought retatrutide online went from 10 mg to 20 mg weekly, injected again the next day after forgetting the first shot, and turned up at hospital with acute kidney injury. He recovered. It is not indexed in PubMed, and the journal would not load for me, so I am reporting it from secondary coverage rather than the paper. (DOI 10.7326/aimcc.2025.1218)

07What nobody knows

What is actually in a vial. I could not find a single peer reviewed chemical analysis of grey market retatrutide. That is a statement about my search, not a claim that none exists.

Why the thing people report most runs backwards. A preprint, not peer reviewed, mapped symptoms in Reddit discussion of grey market use. Of 13,589 people reporting current use, 7,823 had at least one mapped symptom, and the most common one in that group was increased appetite, in a drug whose entire mechanism is appetite suppression. (DOI 10.64898/2026.05.28.26352819) I could not find a published explanation for it.

08My take

The evidence is not the weak point here. The weak points are downstream. It is approved nowhere, so anything anyone holds came through a channel with no assay behind it, and the circulating doses are vial arithmetic rather than trial arithmetic.

Then there is the registry record. A photocopy of real work filed under a new sponsor with a new date, and the only member of its family that does not confess. It copies the escalation schedules for the 4 mg and 8 mg arms, detail you reach for only if you want it to survive being read. Still there, marked recruiting.

If you have a certificate of analysis for the batch you are actually holding, I would like to know what the content line said against the label. One vial came back at 121% of stated. I have no idea whether that is typical.

Frequently asked

Is retatrutide FDA approved?+

No. Lilly's own medical information is unqualified: retatrutide is investigational, approved in no country, and cannot be legally sold or marketed for human use. Lilly says it plans to file in the first quarter of 2027. There is exactly one legal route to it outside a trial, a Lilly expanded access record, NCT07629401, which is physician initiated and case by case.

How much weight did people lose on retatrutide?+

In phase 3, the TRIUMPH-1 master protocol with 2,335 participants, the efficacy estimand gave 19.0%, 25.9% and 28.3% at 80 weeks for 4, 9 and 12 mg against 2.2% on placebo, and the more conservative treatment regimen estimand gave 17.6%, 23.7% and 25.0%. The phase 3 numbers were announced by the company on 21 May 2026 and have not been published in a peer reviewed journal.

What are the side effects of retatrutide?+

In phase 3 at 12 mg against placebo, nausea ran 42.4% versus 14.8% and vomiting 25.3% versus 4.8%. Discontinuation for adverse events climbed across the doses, 4.1%, 6.9% and 11.3%, against 4.9% on placebo, so more than one in nine people quit the top dose over side effects. That is where the tolerability ceiling shows.

Where do the retatrutide doses people use online come from?+

The charts that circulate start at 0.25 to 0.5 mg weekly and aim at 4 mg as a maintenance target, and some split the week across three injections. No published trial used a sub 1 mg maintenance target or split dosing; the phase 2 diabetes 0.5 mg arm is the only published exposure below 1 mg. Phase 2 used 1, 4, 8 and 12 mg as maintenance and phase 3 used 4, 9 and 12 mg, and the registry shows the 12 mg arm started at 2 mg, so the ladder people copy is not what the trial did. The 10 and 20 mg figures come from one hospitalised patient and are not a norm.

What is known about retatrutide bought online?+

I could not find a single peer reviewed chemical analysis of grey market retatrutide, which is a statement about my search rather than a claim that none exists. The only vial level numbers I found were vendor commissioned: a 10 mg vial from one shop came back 99.76% pure but measured 12.05 mg, 121% of the label, with no sterility screen. On harm, one case is the only peer reviewed report I could find, a 32 year old man who bought retatrutide online, went from 10 mg to 20 mg weekly, injected again the next day after forgetting the first shot, and turned up at hospital with acute kidney injury. He recovered.

Is the retatrutide trial that says it is recruiting a real study?+

NCT07467447, sponsor Hudson Biotech, is marked recruiting at a single site, Peking University Shenzhen Hospital, and its official title is character for character identical to Eli Lilly's NCT04881760, a study that completed on 22 November 2022. Nothing in its own text gives it away: I read its complete brief summary and detailed description and neither contains the word fictional, example, mock, simulated, hypothetical or illustrative. It is the one retatrutide registry entry that must never be cited. The cluster it belongs to was found by Morgan McSweeney and published on 31 July 2026.

---

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.