Research-only peptides: what the documented evidence actually shows

The phrase “research-only peptide” is a marketing convention, not a chemical category. The vials sold under that label contain — or are meant to contain — the same molecules sold by FDA-approved drug manufacturers and by 503A and 503B compounding pharmacies. What differs is the supply chain: a website, an offshore manufacturer, a flat disclaimer that the product is “not for human consumption,” and no quality system between the synthesis run and the buyer’s freezer.

A growing peer-reviewed literature, an active FDA recall record, and a substantial anti-doping forensic dataset have, between them, characterized what is actually inside those vials, and what has happened to the people who inject them. This piece surveys that documented evidence. It does not address efficacy. It does not editorialize about the buyers. It reports what cited investigators have found.

What’s actually in the vial: identity and impurity failures

The largest single dataset on the contents of “research-only” peptide products comes from the Cologne anti-doping laboratory’s analysis of 1,190 products seized at the Swiss border (Weber et al. 2017, PMID 28156209). Twelve percent of the seized inventory was labeled as peptide hormones or growth factors. The investigators reported that

For the majority of the investigated products, the pharmaceutical quality was an unsatisfactory standard … less than 20% of the products contained the claimed substance in the respective amount.

A 2018 forensic study in Talanta by Janvier and colleagues profiled the ten most frequently encountered falsified polypeptide drugs on the Belgian market (PMID 30029448). The team characterized API content, related substances, small-molecule contaminants, elemental impurities, and residual solvents in each product class. The authors concluded:

Since these falsified preparations are produced outside the legally required quality systems, end-users have no guarantee regarding the efficacy and safety of these products.

A 2021 Forensic Science International analysis by Fabresse and colleagues examined 75 pharmaceutical products seized from the French black market among bodybuilders (PMID 33838562). The numerator and denominator are quotable: 33% of the products were substandard (out-of-range dosage), 32% were qualitatively counterfeit, and only 19% matched their labeled formulation. The same dataset reported that 80% of seized anabolic-androgenic steroid products were non-original.

For somatropin specifically — the most price-stable peptide hormone in the gray market — a 2017 International Journal of Clinical Pharmacy study by Vida and colleagues test-purchased the molecule from 17 internet vendors (PMID 27888454). Ninety-four percent of the vendor websites required no prescription. Every online sample carried significantly lower somatropin concentration than its label claimed (p < 0.001). The authors noted that “unprofessional distribution and handling is likely to cause degradation and possible patient safety concerns.”

For semaglutide, the most recent systematic survey is the 2024 paired analysis by Ashraf and colleagues in JAMA Network Open (PMID 39093567) and the Journal of Medical Internet Research (PMID 39509151). The team conducted no-prescription test purchases from illegal online pharmacies, ran LC-MS content analysis alongside microbiological sterility and endotoxin assays, and characterized the web-traffic footprint of the seller market. Their summary observation:

All purchased vials were considered probable counterfeit … the top 30 domains directly or indirectly affiliated with illegal online pharmacies accumulated over 4.7 million visits.

Across these five peer-reviewed studies — Belgian, French, Swiss, Hungarian, and the recent JAMA/JMIR work — the labelled identity and labelled dose of an online or seizure-channel peptide product matches the contents in roughly one in five cases. The remaining four contain something else: a different molecule, a sub-labelled dose, an impurity profile that has never been characterized, or a sterile-fill failure that no buyer can detect from the outside of a vial.

Sterility and endotoxin: the FDA recall record

The supply-chain story does not end at the offshore-vendor border. FDA-registered 503A and 503B compounding pharmacies — the licensed US channel — also generate a steady stream of peptide recalls, the bulk of them filed under the agency’s standardized reason “Lack of Assurance of Sterility.” The FDA publishes these actions through its Recalls, Market Withdrawals, & Safety Alerts index and the drug-recall program page.

Counts queried against the openFDA enforcement endpoint on May 2, 2026 illustrate the scale by molecule: 92 recall records mention sermorelin in the product description, 43 mention semaglutide, 21 mention tirzepatide, 6 mention BPC, 5 mention thymosin, and 3 mention melanotan. The dominant recall reasons in those records are “Lack of Assurance of Sterility” and “Lack of Processing Controls.” The named compounding firms include GenoGenix LLC (Boca Raton, FL), Promise Pharmacy LLC (Palm Harbor, FL), North American Custom Laboratories d/b/a FarmaKeio (Richardson, TX), Tailor Made Compounding (TMC Acquisition LLC), Advanced Nutriceuticals LLC, Wells Pharmacy Network LLC, KRS Global Biotechnology, and Pharm D Solutions of Houston. Recalled product lines across these firms have included BPC-157 injectables, tirzepatide lyophilized vials, GHK-Cu copper-peptide injection, thymosin beta-4, melanotan I and II, and combination sermorelin / GHRP-2 kits.

The single openFDA recall record that names a documented patient-harm event involves a sermorelin preparation from Main Street Family Pharmacy of Newbern, Tennessee, dated May 28, 2013. The reason field on that record observes that “the firm received seven reports of adverse reactions in the form of skin abscesses potentially linked to compounded prescriptions.” The Main Street record is unusual in the recall corpus because most sterility recalls are issued prospectively, on the basis of process deficiencies identified during inspection rather than after-the-fact reports of abscesses or sepsis. That a recall reason field references seven separate abscess reports in a single sermorelin lot is the strongest direct compounder-linked adverse-event signal in the public US recall record for peptides.

Investigative reporting has surfaced parallel concerns about active US compounders. A 2025 Endpoints News investigation into Empower Pharmacy — among the largest US compounders distributing GLP-1 and peptide products — reported that

Empower Pharmacy has used poor-quality ingredients, skirted regulations and pushed out or sued employees who raised concerns, an Endpoints News investigation found.

The recall record and the journalism record converge on a narrow point: even within the US-licensed compounding channel, sterility and processing-control failures recur at a frequency that the FDA enforces against in writing every few weeks. The “research-only” channel sits one step further from that enforcement perimeter, with no inspection regime at all.

Documented patient harms: the Melanotan II case-report record

Among the peptides commonly sold under “research-only” labels, Melanotan II has the most developed published case-report literature on direct patient harms. The case-report record spans rhabdomyolysis, renal infarction, priapism, posterior reversible encephalopathy syndrome, and several reports of melanocytic change and melanoma.

The canonical US toxicology case report is Nelson, Bryant, and Aks (2012) in Clinical Toxicology (PMID 23121206). The authors described a 39-year-old man who self-injected 6 mg of Melanotan II purchased over the internet. The patient developed sympathomimetic toxicity, mydriasis, and tachycardia, with a creatine kinase peak of 17,773 IU/L twelve hours later. The report observed:

A 39 year-old Caucasian male injected subcutaneously 6 mg of Melanotan II purchased over the Internet … his CPK elevated to 17773 IU/L 12 hours later.

A 2020 CEN Case Reports paper by Peters and colleagues added renal infarction to the documented event spectrum (PMID 31953620), with the authors discussing thrombotic and direct-toxic mechanisms in the context of melanocortin-receptor-related vascular events. Two case reports of priapism — Devlin, Pomerleau, and Foote (2013) in Clinical Toxicology (PMID 23537392) and Dreyer, Amer, and Fraser (2019) in BMJ Case Reports (PMID 30796078) — described the same melanocortin-pharmacology adverse event in two unrelated patients. A 2013 letter in Annals of Internal Medicine added posterior reversible encephalopathy syndrome (PMID 23648958) — a hypertensive-ischemic CNS syndrome — to the published record.

The cutaneous and oncological end of the case-report literature was consolidated in a 2017 International Journal of Dermatology review by Habbema and colleagues (PMID 28266027). Habbema et al. identified four case reports of melanoma arising in pre-existing moles during or after melanotan exposure:

Four case reports have described melanomas emerging from existing moles either during or shortly after the use of melanotan.

A 2025 case report by Yassin Alsabbagh, Bhujel, and Singh in the International Journal of Oral and Maxillofacial Surgery added a new exposure route and a new tumor location to the record (PMID 40210573): a 22-year-old woman who developed histologically confirmed mucosal malignant melanoma in the anterior maxilla after using a Melanotan II nasal spray for tanning.

No single case report establishes causation. The published series, taken together, characterize a cluster of melanocortin-receptor-related events — vascular, urological, encephalopathic, and pigmentary — that have recurred across continents and across more than a decade of unregulated melanotan use. The case-report literature is the evidence base the FDA placed in front of the Pharmacy Compounding Advisory Committee when Melanotan II was scheduled for review.

Hidden ingredients and novel analogues: WADA’s accidental product-quality lab

The world’s most systematic public-product-quality testing infrastructure for peptides is, somewhat by accident, the international anti-doping system. WADA-accredited laboratories receive seized vials, athlete samples, and over-the-counter supplements as a routine matter, and they publish what they find.

A 2019 Drug Testing and Analysis paper by Popławska and Błażewicz at the Medical University of Lublin de-novo-sequenced a previously unreported heptapeptide from a seized black-market injection vial (PMID 30051972). The compound — molecular weight 874.02 Da, identified as a glycine analogue of GHRP-2 — appeared in no reference database. The authors observed:

The results of this study may indicate a new approach to circumvent a detection of doping practices.

The implication is that a vial sold to a buyer as “GHRP-2” may contain a related but legally and pharmacologically distinct molecule — a novel chemical entity that has never been characterized in animals, never tested for receptor selectivity, and never assessed for impurity or aggregation behavior. The buyer cannot tell from the label.

The reverse pattern is also documented. A 2010 Drug Testing and Analysis paper by Thomas and colleagues at the Cologne anti-doping laboratory identified and quantified approximately 50 µg of GHRP-2 per tablet in a product marketed as an over-the-counter nutritional supplement (PMID 20878896). A consumer purchasing what was marketed as a benign supplement received a WADA-prohibited peptide hormone secretagogue.

A related finding came from Cox and Eichner (2013) at the Sports Medicine Research and Testing Laboratory, who documented human insulin-like growth factor-1 in commercial deer antler velvet supplements marketed as natural performance products (PMID 23996390). The Weber 2017 Swiss-border dataset discussed above (PMID 28156209) sits in the same evidence category, since it was conducted by a WADA-accredited group on customs-seized inventory.

What the doping literature contributes that the clinical literature cannot is a continuous testing pipeline: a stream of seized vials and unlabeled supplements that arrive at accredited laboratories and are subjected to identity, content, and impurity analysis. The published findings have, on multiple occasions, identified peptide ingredients that the seller’s label did not disclose and chemical entities that the public scientific record had not previously seen.

Injection-population harms: bloodborne viruses and abscesses among IPED injectors

Direct case-report literature on sepsis specifically attributable to research-grade injectable peptide use is sparse on PubMed. The closest population-level evidence comes from the United Kingdom’s surveillance of men who inject image- and performance-enhancing drugs (IPEDs) — a population that overlaps but is not identical to “research peptide” buyers.

A 2016 Journal of Acquired Immune Deficiency Syndromes analysis by Hope and colleagues at Public Health England drew on cross-sectional prevalence surveys in England and Wales between 1992 and 2013 (PMID 26361173). The team observed:

Blood-borne virus prevalences among IPED injectors have increased and for HIV, is now similar to that among psychoactive drug injectors.

A companion 2015 Epidemiology and Infection survey by Hope and colleagues of 366 male IPED injectors quantified the localized infectious burden (PMID 24713416). The investigators reported:

Of the 366 male IPED injectors surveyed, 42% reported ever having redness, swelling and tenderness … 6·8% had ever had an abscess or open wound at an injection site.

The IPED-injector category in the Hope surveys is broader than peptide-only injection: it captures anabolic-androgenic steroids, growth-hormone analogues, peptide hormone secretagogues, and the various adjuvants those populations co-administer. This population overlaps but is not identical to “research peptide” buyers. The inferential limit is real and worth stating in plain terms: these data describe an injecting population in which gray-market peptide products are one component, not a peptide-specific cohort. The signals — rising bloodborne-virus prevalence and a measurable lifetime abscess rate — are nonetheless the strongest direct population-level evidence currently available on injection harms in the population that buys “research peptides.”

What this evidence does NOT support

For BPC-157, GHK-Cu, ipamorelin, MK-677, and IGF-1 LR3, the published case-report literature on harms attributable to research-grade gray-market product is essentially absent. This is not evidence of safety — it is evidence of what the literature has and has not investigated. The harm story for those molecules rests on the supply-chain quality data — Weber, Janvier, Fabresse, Vida, Ashraf — and on the FDA recall record, not on patient case-reports. A reader weighing those substances against the evidence in this piece should treat the absence of case-report literature as a gap, not a clearance.

Closing

The documented evidence on research-only peptides is, at present, evidence about a supply chain. It is most quantitative on identity and dose (less than 20% of seized products match label, in the largest dataset), on sterility (more than a hundred FDA recalls of compounded peptide products in seven years), and on a single peptide — Melanotan II — for which the case-report literature spans rhabdomyolysis, renal infarction, priapism, encephalopathic events, and melanoma. It is thinner on direct patient outcomes for most other peptides, and the IPED-injector population studies stop short of peptide-specific attribution. The published record does not, however, support a default assumption that an internet-sourced peptide vial contains what its label claims, in the dose its label claims, free of unacceptable bioburden.

---

Methodology

The evidence surveyed in this piece is drawn from a curated source pool of forty verified items compiled from PubMed-indexed peer-reviewed literature, the FDA recall and warning-letter record, the openFDA enforcement endpoint (api.fda.gov/drug/enforcement.json), and whitelisted journalism (Endpoints News, Reuters). Source identifiers (PMIDs, recall dates, openFDA queries) and direct quotations are reproduced from that pool. Every cited claim links to a source document on the pepi.ninja whitelist. The full source compilation underlying this piece is documented internally as docs/research/research-only-peptide-harms-sources.md and was assembled on May 2, 2026.