Does research retatrutide contain retatrutide? The lab record

Answer first: in the certificates vendors publish, research retatrutide almost always contains retatrutide. Across five public certificate sources we mirror, 319 of 322 results with an identity test confirmed it. Three failed, and one vial held nothing measurable. The median vial held 110.2% of its label. That record is vendor-selected. Finnrick, which says it publishes results whether or not the vendor likes them, records 31 of 194 recent retatrutide tests as failed identity.

What set this off

Three pieces in one week put the question in front of a general audience.

• Women's Health UK, 11 October 2026. Abby Hepworth quoted Holly Lofton, MD, director of medical weight management at NYU Langone and a retatrutide trial investigator. Lofton said patients who took grey-market reta reported no weight loss, and that "these compounds often don't contain any active ingredient." The piece cites no lab results for that line.
• JAMA Medical News, 9 October 2026. Rita Rubin reported two analyses from Australia. Australia's Therapeutic Goods Administration said a product sold as retatrutide contained a high dose of semaglutide and no retatrutide. University of Queensland researchers tested three products labelled 10 mg: all three contained retatrutide, one at half the label and two above it.
• BMJ Case Reports, 7 October 2026. Raphael Buzzi and Lars Huber of University Hospital Zurich describe a woman in her mid-20s admitted with refractory nausea after four weekly injections of an unverified product reported to contain retatrutide, given in a wellness setting by a life coach with no medical qualification. The authors score the link as probable (Naranjo score 7), not proven. The product was not analysed, so the case says nothing about what the vial held.

The claim is that the active ingredient is often missing, and the evidence offered for it is clinical anecdote. Lab records can test it.

What the published lab record shows

We pulled every retatrutide record from our mirrors of Kovera Labs (to 2 August 2026), Janoshik's public listing (30 June 2026 snapshot) and the certificate pages of Particle Peptides, Ascension Peptides and Injectify. We matched code-named products to retatrutide where the certificate or lab listing names the compound, and left out retatrutide/cagrilintide blends. Where the same test appeared under more than one client name, we counted it once. These are the same 323 results as the retatrutide row of our purity and fill study.

Test dates run from 4 February 2025 to 16 July 2026, across 130 client names. Janoshik's 67 records drop to 61 because two results were issued under several test numbers: one 22.39 mg result on tests 137804 and 147610 to 147614 (six numbers, 14 April 2026), and one 11.01 mg result on tests 71592 and 71881 (19 July 2025). Each counts once.

Identity. Three results of 322 failed, a rate of 0.9%. All three are Kovera certificates you can open by report number in the Kovera index:

• KVR-2026-06D5D6, analysed 22 May 2026: labelled retatrutide 10 mg. Measured 0.00 mg, purity 0.000%, identity failed. This is the only record in 323 that matches the "no active ingredient" description.
• KVR-2026-9197F8, client printed as Instant Peptides, analysed 11 May 2026: sold as "GLP-3 RT". Purity 99.85%, LC-MS identity "Unknown", fail.
• KVR-2026-40AF14, client printed as Glacier Aminos, analysed 11 May 2026: sold as "GLA-3 RT". Purity 99.88%, identity "Unknown", fail.

The last two held peptide. Each read about 99.9% pure, so the lab found a clean peptide and could not match it to the specification, which the certificate lists by the code name. The record does not say what the peptide was. Read them as "identity not confirmed". They do not prove a substitute.

Content. Besides the empty vial, 307 results report a measured mass. Overfill is the norm:

The middle half of vials held between 106.2% and 115.1% of label. The two under 90% are Janoshik test 102150 (8.78 mg in a 10 mg vial, 26 February 2026) and Particle Peptides lot PO-2026302 at 8.45 mg on 2 March 2026. The same Particle lot measured 9.95 mg when retested on 14 April 2026. One result describes one vial.

Purity. 320 results report it. The median is 99.86%, the lowest is 98.67%, and 12 results sit below 99%.

How many results were sold under code names?

106 of the 323 results (33%) were sold under a code name or abbreviation rather than "retatrutide". We counted 36 distinct strings. The most common:

Others include ZEN-3, R3, RT, R-10, R-30, TriVial-3R, BioEnhanced-3R and "GLP/GIP/GLUCAGON". Two of the three identity failures were code-named; one of 217 named retatrutide results failed against two of 106 code-named ones. Three events cannot carry a comparison. Our GLP-3 RT explainer covers why sellers rename the molecule and which names FDA's letters treat as retatrutide.

The other record: Finnrick

Every certificate above exists because a vendor paid for it and agreed to publish it. A vendor with a bad batch can leave it unpublished. That is why the 0.9% figure is a floor, not an estimate.

Finnrick works differently. Its methodology page says it sources samples from vendors and sends them to third-party labs, and it also runs free community testing for US individuals. Results are published whether or not the vendor likes them. On 12 October 2026 its public retatrutide page listed 4,367 tests across 378 vendors, from 17 December 2024 to 25 September 2026.

The page shows the 200 most recent tests (15 to 25 September 2026): 131 pass, 63 fail, 6 incomplete. The visible table gives only pass or fail. The data the page serves to any visitor, without a login, also carries a reason code for each failure, which we counted:

That is 31 identity failures in 194 classified tests, 16.0%, by Finnrick's reason codes as read on 12 October 2026, spread across 24 of the 114 vendors in the window. Finnrick grades with a stricter purity bar (99.5%) than most certificates use, so its purity fails are not comparable to the figures above. Its identity fails are comparable: the lab did not find the expected compound. We did not open the underlying certificates, which Finnrick keeps behind a membership, so we cannot say how many of the 31 held another peptide and how many held nothing.

Reading the two records together

Each record answers a different question.

• The published certificate record says that when a vendor sends retatrutide to Kovera or Janoshik and publishes the result, the vial almost always holds retatrutide, usually more than the label.
• That record cannot rule out vials nobody tested, batches a vendor chose not to publish, or a seller whose shop page shows a certificate for a different batch. "Most tested vials contained retatrutide" does not prove most vials sold do.
• Finnrick's recent tests, which a vendor cannot hold back, found the wrong or no compound in about one in six. That is far from "often no active ingredient" as a description of the market, and far from 0.9%.

Michael DiStefano, a University of Colorado researcher JAMA quoted, asked: "Maybe it is actually retatrutide, but is it the right potency?" The lab record says potency usually runs high. Our inference, not a finding: identity failures sit mostly in products whose sellers do not publish certificates.

How to check the lot in your hands

1. Find the report number on the certificate the seller gave you. Kovera uses KVR-2026-XXXXXX. Janoshik uses a task number plus a unique key.
2. Resolve it at the lab. Paste the code into our COA verifier, which checks it against the Kovera and Janoshik records we mirror, then open the record on the lab's own site. Our Janoshik verification walkthrough shows the four checks for that lab.
3. Match the batch. The batch or lot on the record must match the code on your vial. A certificate for another batch, or a generic batch like "2026", tells you about a different vial.
4. Read identity, not just purity. A 99.9% result with a failed or "Unknown" identity line is what two of the three failures above looked like.
5. No certificate for your batch? Then the only answer is a test of that vial. Our labs comparison covers price and turnaround.

If retatrutide seems to have stopped working, the troubleshooting guide separates a bad vial from storage and timing. The Janoshik retatrutide and tirzepatide results break down that lab's 68 reports by vendor.

Method

Mirrors: Kovera Labs public verifier (records to 2 August 2026), Janoshik public listing (30 June 2026), Particle Peptides, Ascension Peptides and Injectify certificate pages. A record counts as retatrutide when the lab listing or certificate names retatrutide, or the product carries a code name the lab listing maps to it. Blends with cagrilintide are excluded. Duplicates are records from the same source with the same measured mass, purity and test date, which catches one test reissued under several client names; 348 records became 323. This is the same rule our purity and fill study uses. Content is measured mass divided by label mass. Finnrick totals and pass/fail counts are as shown on its public retatrutide page on 12 October 2026; the failure reasons come from the data that page serves without a login.