It Was Sold as Retatrutide. The Regulator’s Lab Found No Retatrutide at All

On 14 September 2026 the Therapeutic Goods Administration published the result of its own laboratory analysis of a vial sold as retatrutide, after a consumer who used it vomited uncontrollably, tore their oesophagus and was treated in hospital.¹ The vial contained no retatrutide. It contained undeclared semaglutide, at roughly eight times the level found in semaglutide products the TGA has assessed and entered in the Australian Register of Therapeutic Goods. The case is worth reading closely because it fails at a step most discussions of peptide quality never reach: not how pure the contents were, but whether the contents were the declared molecule at all. Condor Research supplies reference materials for laboratory research use only; this article reports a regulatory finding and contains no human use guidance.

What the TGA actually found

The alert is short and unusually specific. The TGA received a report of a serious adverse effect linked to a product sold as retatrutide. During its investigation the regulator obtained a sample of the product in its original unmarked vial and tested it in its own laboratory. Two findings are reported: the product contained undeclared semaglutide and no retatrutide, and the amount of semaglutide detected was approximately eight times higher than the level found in registered semaglutide products.¹

Retatrutide is an investigational triple hormone receptor agonist in Phase 3 development at Eli Lilly,⁴ and Lilly states plainly that it is legally available only to participants in its clinical trials.² The TGA makes the same point from the regulator’s side: retatrutide has not been approved by the TGA or any comparable international regulator, so any retatrutide product obtained outside a trial could be counterfeit.¹

Professor Robyn Langham, the TGA’s Chief Medical Adviser, framed the risk in a sentence that is, in effect, a list of separate analytical questions: “The biggest risks with unapproved peptide products are not knowing what is in the vial, how much of or even what substance it contains, whether it is sterile, or whether it contains other contaminants or toxins.”¹

Four questions, four different tests. A single number on a certificate answers at most one of them.

Why a purity figure would not have caught this

The reflex response to a case like this is to ask for a purity result. It would not have helped, and understanding why is the transferable part of the story.

Chromatographic purity by HPLC is a relative measure. The instrument separates what is in the sample into peaks, integrates the area under them, and reports the main peak as a percentage of the total. It describes homogeneity: how much of what is in the vial is one substance rather than several. It does not, by itself, say which substance that is. A vial of semaglutide analysed against a semaglutide reference can return an excellent purity figure and remain, from the buyer’s point of view, entirely mislabelled. The number was never answering the question “is this retatrutide”.

Identity is established by comparing the sample against a known reference: intact mass by mass spectrometry, fragmentation by LC-MS/MS, or amino-acid sequencing. Semaglutide and retatrutide are different molecules with different masses and different sequences, and either technique distinguishes them immediately. The distinction only holds, however, if the laboratory is asked to identify the contents rather than to quantify a compound it has been told is present. An assay set up to measure retatrutide against a retatrutide reference can fail to see semaglutide at all.

Content, the third question, is separate again. Quantity requires a calibration curve against a reference standard of known concentration, and it is the question that carries the actual consequence in this case. The TGA did not merely find the wrong molecule; it found roughly eight times the level present in assessed products.¹

What each test answers, and what it cannot

Read down the middle column and the TGA’s sentence resolves into its parts. We have taken each of these apart at length elsewhere: how to read a certificate of analysis, what “99% pure” on a COA really means, and the endotoxin and sterility section almost nobody reads. What is in the vial is identity. How much is content. Whether it is sterile is the sterility test. Other contaminants or toxins covers endotoxin, residual solvents and elemental impurities, each with its own method. A certificate that reports one of these is not a weaker version of a certificate that reports all of them. It is an answer to a different question.

How often does the declared contents turn out to be wrong?

The Australian case is a single adverse event, and single cases invite the objection that they are outliers. There is now a larger dataset to put next to it. Mendias and Awan analysed 6,441 peptide samples across fourteen compounds, including BPC-157, retatrutide, semaglutide, tirzepatide, CJC-1295, GHK-Cu, ipamorelin, PT-141, TB-500 and others, drawn from a large publicly available independent testing dataset of gray-market preparations sold directly to consumers.³

Applying two acceptance frameworks, one approximating the standards for 503A compounded medications and one closer to the standards applied to FDA-approved peptide drugs, they reported that between 41.6% and 71.1% of samples failed to meet basic quality criteria, and that measurable endotoxin contamination was present in 15% of samples.³ The preprint is not peer-reviewed, and its dataset is self-selected: it reflects material that someone chose to submit for testing, which if anything biases towards the more careful end of the market.

The scale of the market those samples come from is itself now measurable: documented use of BPC-157 alone rose 33-fold between 2020 and 2026, and roughly a third of the users whose source was recorded named a gray-market vendor, which we cover in what the first real-world data on BPC-157 actually shows.

Their closing observation is the one worth keeping. Consumer-directed third-party testing improves transparency, they write, but captures only a small fraction of the safety profile relevant to injectable peptide preparations.³ A published test result is evidence about the sample that was tested, not a property of the vial in front of you.

What this means for laboratory material

The grade a material is sold under is not a formality either: one molecule can exist in three separate legal categories with three different sets of obligations, and establishing that two peptide preparations are the same thing is harder than it sounds.

For a research laboratory the practical consequence is narrower and more mundane than the news story suggests, and it has nothing to do with self-administration. If the identity of a reference material is wrong, every experiment run with it is measuring something else. A dose-response curve built on a mislabelled compound is not a weaker result; it is a result about a different molecule, and it will not reproduce.

The questions that follow from the TGA finding are the ones any purchaser of reference material can ask, in ascending order of what they establish: which compound was analysed, against which reference standard, by which method, on which batch, and on what date. A certificate that names a method and a batch is checkable. One that reports a percentage with no method, no reference and no batch is not a weaker certificate. It is an unverifiable statement.

Readers who want the analytical background rather than the news will find it in our note on what HPLC and mass spectrometry can establish about a peptide sample, in our review of the evidence that TFA is not an inert counterion, and in our companion piece on what endotoxin testing measures and chemical purity cannot. The regulatory context is covered in Research Use Only is not a regulatory shield, in why an RUO label is not a legal status, and in the EMA’s first synthetic-peptide guideline.

How this was checked. The TGA safety alert was read directly on tga.gov.au on 21 September 2026, and the quotations from Professor Langham are taken from that page rather than from press coverage. Retatrutide’s development status was taken from Eli Lilly’s own EASD 2026 press release of 15 September 2026. The gray-market testing figures are from a preprint, and are reported here as a preprint: not peer-reviewed, self-selected sample, authors’ own acceptance thresholds. Method descriptions are general analytical chemistry and are stated without reference to any particular supplier’s practice. Version 1.0, first published 21 September 2026.

Condor Research supplies peptides as characterised reference materials for laboratory research use only: not for human or veterinary use, not for diagnostic or therapeutic application, and not for any food or cosmetic purpose. Nothing in this article describes, recommends or implies human use of any compound, and nothing in the Australian case should be read as a protocol, a warning about dosing, or advice about self-administration, which falls entirely outside the scope of research material supply.

Condor Research · Scientific desk
Atrio Sciences s.r.o., IČO 57 669 651, Nitra (SK) · info@condorresearch.com