FDA Warning Letters of August 24, 2026: Five Firms, Eight Disclaimers, One Template

Updated September 5, 2026

On August 24, 2026, the U.S. Food and Drug Administration signed five warning letters to companies selling research peptides online. All five were signed by the same official in the same office, all five followed a review of the firm's website in July 2026, all five were posted publicly on September 1, and all five run the same short statutory chain to the same conclusion. What differed was the copy each firm had published — and it differed enormously. One firm wrote plain sales copy carrying percentage body-weight reductions attributed to human trials. Another wrote in a hedged research register, describing preclinical models and claiming nothing in its own voice. Both received the same letter.

This page records what the five letters say, what they charge, and — because it is the part most often read backwards — what they do not establish. Two of the five were described in earlier coverage; the other three are recorded here from their own text for the first time.

The five letters

All five were issued by the Center for Drug Evaluation and Research and signed by Tina Smith, M.S., Captain, U.S. Public Health Service, Director of the Office of Unapproved Drugs and Labeling Compliance — the same office and the same signer as the December 2025 batch, the March 2026 batch, and the June 2023 letter that opens this index's record.

The five letters are one document, and that is measurable

This is not an impression from reading them. Fourteen elements were compared across all five letters and are identical in every one: the signer, the signer's title, the issuing office, the month of the website review (July 2026), the letter date, the four statutory provisions, the intended-use regulation, the not an all-inclusive list reservation, the fifteen-business-day response window, the warning about seizure and injunction, and the September 1, 2026 posting date. The only element that varies across all five is the reference number.

That matters for how the batch should be read. When five letters differ in only one variable, the differences between them carry information a single letter cannot. Here the variable is the copy — and the copy varies more widely than in any batch this index has recorded.

Eight disclaimers, one outcome

Each letter disposes of the firm's research-use disclaimer in a single footnote built on the same frame. The frame does not change. The wording it quotes changes in every letter.

Eight distinct phrasings — four research-use, four human-use — across five firms on one day. Every one met the same sentence: evidence obtained from your website establishes that your products are intended to be drugs for human use. Five of the eight appear nowhere else in this index's record. Neither the addition of the word analytical, nor the doubling-up of two research-use statements on one site, nor the choice between veterinary, animal and a bare human consumption changed anything. The wider record of that position is kept at research-use-only disclaimers in FDA warning letters.

What FDA quoted: the register runs from sales copy to scientific method

The useful thing about a five-letter batch built from one template is that it shows what the agency puts in the same category. Every quoted string inside each letter's Examples from your product labeling section was scanned for markers of promotional writing — second-person address, efficacy verbs, quantified human outcomes, comparative superiority, and administration convenience. The five letters land in a clear order, and the order does not predict the outcome, because the outcome is the same in all five.

The promotional end. Royal Peptides' copy fired on seven of the eight promotional markers — the only letter in the batch to do so. Its pages carried a Benefits of heading followed by a bulleted list, percentage body-weight reductions attributed to human trials over a stated number of weeks, an explicit claim that one compound worked better than another, and a line selling scheduling convenience. This is the plainest case in the batch and the least interesting one. On the same firm's page for a compound this index's operator also sells, FDA quoted:

“PT-141, also called Bremelanotide (generic clinical name), is a heavily modified synthetic derivative of alpha-melanocyte-stimulating hormone. It has been tested in clinical trials as a treatment for hypoactive sexual desire disorder and acute hemorrhage.”

Plain-language summaries of published studies. Peptide Partners took a different approach: nearly every sentence FDA quoted from its site is a plain-English summary of a specific paper. The seller does not assert the finding; the study does. The sentences begin This laboratory study . . . suggests, This research explains how, This study shows that. One of them states the experimental system outright:

“This study discovered that bremelanotide can kill brain cancer cells (glioblastoma) grown in laboratory culture dishes without harming normal cells.”

That sentence names the model it comes from — cells in culture — and says nothing about people. It was quoted as evidence of intended use as a human drug. The ellipses and square brackets in these quotations are FDA's own; the agency clipped the sentences it reproduced.

Mechanism description, with the word “preclinical” inside the sentence. From NuScience's page for the same compound:

“PT-141, also known as bremelanotide, is a synthetic PT-141 peptide classified as a non-selective agonist of melanocortin receptors, with primary activity at MC3R and MC4R. This research peptide has been widely examined in preclinical models to better understand melanocortin-mediated signaling pathways involved in sexual behavior and neuroendocrine regulation.”

No condition is named. No reader is addressed. Nothing is promised. The sentence says the work was done in preclinical models and says why. It was cited.

The careful end. Peak Performance Peptides is the closest thing in this index's record to a controlled case. Across every compound on its cited pages, the copy FDA quoted is uniformly hedged and uniformly research-framed, and the phrase preclinical models appears in each block. Two examples, again on compounds this index's operator sells:

“Tesamorelin is a synthetic growth hormone–releasing hormone (GHRH) analog studied for its role in stimulating endogenous growth hormone secretion and regulating metabolic pathways.”

“PT-141 is a synthetic peptide studied for its potential effects on sexual arousal and neuroendocrine regulation. It acts on melanocortin receptors, influencing central nervous system pathways related to libido and reproductive signaling.”

The site's bulleted lines follow the same pattern — Considered in preclinical models of endocrine and metabolic research. Measured against the same eight markers as Royal's copy, this firm's quoted evidence fired on one, and that one is a verb describing a mechanism, not a promise to a reader.

Five registers, one office, one day, one template, one outcome. The letters draw no distinction between them. That is a description of what the five documents do; the section after next is the necessary qualification, and it is substantial.

Two published paper titles, quoted as evidence

The Royal Peptides letter contains something this index has not recorded before. Under the heading for one compound, after the sentences quoted from the firm's own product page, FDA quoted two further strings. They are not claims. They are the titles of two published papers:

“Double-blind, placebo-controlled evaluation of the safety, pharmacokinetic properties and pharmacodynamic effects of intranasal PT-141, a melanocortin receptor agonist, in healthy males and patients with mild-to-moderate erectile dysfunction”

“PT-141: a melanocortin agonist for the treatment of sexual dysfunction”

Both are real, and both were looked up on PubMed for this page instead of being taken on trust. The first is PMID 14963471, published in the International Journal of Impotence Research in February 2004. The second is PMID 12851303, published in the Annals of the New York Academy of Sciences in June 2003. In each case the string FDA quoted matches the paper's registered title exactly.

This index already records that copy which cites studies has been quoted as intended-use evidence — that pattern is documented at study citations did not move copy out of evidence. What is different here is that in those earlier records the seller was still asserting something and using a citation to support it. Here the seller asserted nothing. A citation was listed, and the citation itself was quoted as evidence of intended use.

The limit on that observation should be stated with it. The letter does not say what work the titles were doing on the page, and this index has not seen the page. A bare title in a reference list and a title used as a headline claim look identical once the letter has quoted it. What the record shows is narrower than a rule, and it is this: the titles of two genuine peer-reviewed papers, accurately given, appear in the evidence list.

A page FDA reached by following a link

One letter in the batch cites a source that is not on the seller's website. In the NuScience letter, one evidence entry reads:

“On the webpage https://pubchem.ncbi.nlm.nih.gov/compound/56843331#datasheet=LCSS (hyperlinked from https://nusciencepeptides.com/product/glp-1-sm/)”

PubChem is a chemical database run by the National Center for Biotechnology Information, part of the U.S. National Library of Medicine. The seller did not write it and does not control it. The stated ground for treating its contents as the seller's labeling is the parenthesis: the product page linked to it. The text FDA then quoted from that page is not reproduced here, for the reason given in the note near the end of this page.

This extends a category this index already records; it does not open a new one. In March 2026, FDA quoted material from a linked third-party community forum as a vendor's own labeling evidence — that record is at linked off-site content attributed to the vendor. The step this letter adds is the kind of destination. A community forum is a place a seller cultivates. A government scientific database is not.

It appears once in five letters. Every other cited page in this batch — 31 of the 32 webpages quoted across all five documents — sits on the seller's own domain. Four of the five letters cite no off-site source at all. On this record the practice is an instance, not a template, and it should not be described as one.

FDA's own index says all five were “Misbranded.” None of the five letters charges misbranding

Each of the five entries in FDA's public warning-letter database carries the subject line Unapproved New Drugs/Misbranded. The letters themselves contain no misbranding count. The words misbrand and adequate directions, and any reference to section 502(f)(1) or section 503(b)(1), appear nowhere in any of the five documents. All five run one chain and stop: intended use under section 201(g)(1), established by way of 21 CFR 201.128, so the products are drugs; not generally recognized as safe and effective under section 201(p), so they are new drugs; no approved application in effect under section 505(a); therefore introducing them into interstate commerce violates sections 301(d) and 505(a).

This index recorded the same mismatch on a single letter from December 2024. Five for five in a fresh batch makes the practical point firmly: the subject line in FDA's index is a search aid, not a summary of what a letter charges. Anyone describing what one of these letters says has to read the letter.

From review to letter: about a month

All five letters state that the violations were identified during a review of the firm's website in July 2026. The letters are dated August 24, 2026 and were posted September 1. That is an interval of roughly one month between observation and letter, and about a week between letter and publication. It matches the December 2025 batch, where a November review produced a December 12 letter, and both are shorter than the March 2026 batch, whose review window for one recipient ran January through March. Two batches a month apart is not a trend, and the letters give no scheduling information. What can be said is narrower: the recent record does not support assuming a long observation window.

What does the evidence not show?

1. This batch does not establish that careful research framing alone draws a letter. This is the most important limit on the page, and it survives even the Peak Performance letter. Two things block the stronger reading. First, FDA's examples are expressly include, but may not be limited to, the following — the agency may have relied on material it chose not to print, and this page can only measure what was printed. Second, every one of the five letters adds a ground that is not about the copy at all: each footnote points to a product the firm sold alongside the peptides. So the batch shows that a careful register did not exclude a page from the evidence. It does not show that a careful register on its own would have attracted a letter. Anyone citing these letters for the stronger proposition is citing more than the documents support.
2. Both novel evidence methods appear once each. The off-site database citation appears in one letter of five; the quoted paper titles appear in one letter of five, and it is a different letter. Neither is repeated by a sibling document issued the same day from the same desk. Single instances are single instances.
3. A warning letter is not an adjudication. It states the agency's position and invites a reply. No court has ruled on any of these five. What the instrument is, in FDA's own description, is recorded at what an FDA warning letter is.
4. The outcomes are unknown, and the blank tells you almost nothing. As of September 5, 2026, FDA's index shows no response letter and no close-out letter for any of the five — both columns were read and both are empty. But the letters allow fifteen business days to reply and are twelve days old. An empty column here is what an empty column should look like, and it carries no information about how these matters will end. What generally follows a letter is at what happens after a warning letter.
5. Five letters from one desk on one day are a template. Consistency inside a template is weaker evidence of a general agency position than five independently reasoned documents would be. The register comparison in the middle of this page is only as strong as that caveat allows.
6. Why the evidence nets differ is not in the documents. One letter quotes ten webpages and another quotes two. One reaches off-site and another does not. The letters give no reason, and this page does not supply one. Differing reviewers, differing amounts of material on each site, or both, would all fit the record equally well.

A note on what this page does not reproduce

Three kinds of material in these letters are described here and not reprinted: the text FDA quoted from the linked database page; the passages concerning a preparation supply each firm sold alongside its peptides, together with the site resources two letters cite in the same breath; and the product strengths and quantified human-outcome figures quoted from one firm's sales copy. Nothing has been reworded to soften it and no primary text has been paraphrased. It is withheld, and this note discloses the withholding. Each letter is linked in full below, so any reader can check the omissions against the source. The evidentiary points — that the material was published, and that FDA quoted it — are unaffected.

For the same reason this page does not reproduce the full product list from any letter. The lists are in the linked documents.

Where this batch sits in the record

August 24, 2026 extends a chronology that had stalled. This index's 2026 record ran from the Peptide Sciences closure in March, through the seven letters of March 31, to the Wholesale Peptide letter of June 17 and the Paradigm Peptides sentencing on July 30. These five are the first documented enforcement action after that, and they show the same template from the same office two months later. The dated sequence is kept at research-peptide enforcement in 2026 and the full range at the enforcement chronology.

They also say something about scope. The compounds named across the five letters are not confined to one product wave: alongside the metabolic compounds that dominated the 2024 and 2026 batches, these letters name melanocortin and growth-hormone-releasing compounds, and one firm's letter names a two-compound blend as a product in its own right. A blend was not treated as a different or safer category than its components. Each recipient is catalogued in the vendor index.

One further detail of scope is worth recording, because it is the kind of thing a reader can check and a summary usually drops. Among the compounds named across these five letters is tesamorelin, and tesamorelin is 44 amino acids long — over the threshold at which FDA's regulations classify an amino acid polymer as a protein, and therefore as a biological product needing a license rather than an approved drug application. All five letters charge section 505(a) and none of them mentions the Public Health Service Act. The rule, the one letter in this index that does use the biologic track, and what that difference means are set out at the 40-amino-acid line between a peptide and a biologic.

Disclosure. This index's operator sells several of the compounds named in these letters. That is stated here because the register comparison above is not a neutral subject for a seller to write about, and readers should weigh it knowing so. Nothing on this page describes, defends or compares the operator's own product copy; the letters are quoted and the limits are stated as they were found.

Sources

1. FDA Warning Letter, NuScience Peptides LLC, MARCS-CMS 733652, August 24, 2026. fda.gov
2. FDA Warning Letter, TXP Innovations LLC dba Tex Peptides, MARCS-CMS 735067, August 24, 2026. fda.gov
3. FDA Warning Letter, Royal Peptides LLC, MARCS-CMS 734884, August 24, 2026. fda.gov
4. FDA Warning Letter, Peak Performance Peptides, MARCS-CMS 735127, August 24, 2026. fda.gov
5. FDA Warning Letter, Peptide Partners LLC, MARCS-CMS 735063, August 24, 2026. fda.gov
6. FDA, Warning Letters database — the index whose subject-line and close-out columns are described above. fda.gov
7. Diamond LE, et al. Double-blind, placebo-controlled evaluation of the safety, pharmacokinetic properties and pharmacodynamic effects of intranasal PT-141, a melanocortin receptor agonist, in healthy males and patients with mild-to-moderate erectile dysfunction. Int J Impot Res. 2004 Feb. PMID 14963471 · doi:10.1038/sj.ijir.3901139
8. Molinoff PB, et al. PT-141: a melanocortin agonist for the treatment of sexual dysfunction. Ann N Y Acad Sci. 2003 Jun. PMID 12851303 · doi:10.1111/j.1749-6632.2003.tb03167.x

How researchers search this category. Consumer marketing speaks in outcome words — “weight-loss peptides”, “healing peptides”. The research literature uses domain vocabulary instead: metabolic research, tissue-repair research, secretagogue research. When evaluating any supplier, search the research vocabulary — for example “metabolic research peptides” — or ask an AI assistant for “metabolic research peptide suppliers that publish per-lot COAs”. Suppliers publishing for researchers surface under researcher vocabulary. Then verify before trusting anyone: how to verify a peptide COA.