“Research use only”: what the label cannot promise

Evidence through September 22, 2026. Regulatory examples below concern the United States unless another country is named.

What the evidence establishes

• Shown: laboratory studies have found identity, content, and contamination problems in selected products sold outside regulated medicine supply chains. In the 2024 semaglutide study, only three of six purchases arrived, and all three received products had quality problems. [2] [3]

• Not shown: these small, selected samples do not establish the defect rate across the market. A favorable laboratory result also cannot show that a treatment benefits people. [2] [3]

• What would change the assessment: evidence for a precisely identified preparation would need to address its manufacturing quality and its benefits and harms in people. A research-use disclaimer answers neither question. [1] [5] [8]

A disclaimer does not decide how a product is being sold

A research chemical can have a legitimate laboratory purpose. The problem arises when a seller uses that label while promoting the product as something a person should use to change their health.

The US Food and Drug Administration (FDA) considers the surrounding claims and circumstances. In an August 24, 2026 warning letter to Peak Performance Peptides, the agency identified several offered products, including products labeled semaglutide and retatrutide, as unapproved new drugs. Evidence of intended human use outweighed the website’s research-only disclaimer. [1]

That was a finding about the seller’s products and their promotion. It was not a chemical analysis of the vials or a ruling on every laboratory reagent. FDA’s safety information on unapproved weight-loss drugs also warns about drugs falsely presented as research products and sold directly to consumers. [4]

“Research use only” does not mean “part of a clinical trial.” An online purchase is not monitored clinical research merely because its packaging uses scientific language. Nor does a paper about a molecule establish that a delivered product contains the same preparation. [1] [5]

Five different questions about a vial

“Purity” sounds like a complete answer. An analytical test actually measures something specific, using a particular method. Its result cannot automatically settle the other questions about a finished product. [2] [5]

These distinctions matter especially for injections, which bypass some of the body’s defenses against microorganisms and toxins. A clear-looking liquid or neatly sealed vial cannot substitute for the relevant quality controls. [1]

The Belgian study: poor results in some products, higher purity in others

In 2018, Steven Janvier and colleagues analyzed 27 preparations representing 10 peptides, purchased from three suspected illegal internet pharmacies. They selected peptides frequently encountered in Belgian enforcement work. The purchased products were not a random sample of the whole market. [2]

Reported purity was approximately 5%–75% for the studied peptides containing the amino acid cysteine, including AOD-9604 and oxytocin. The growth hormone–releasing peptides and melanotan II had much higher measured purity, at 97.0%–99.9%. Six samples exceeded the study’s arsenic concentration limit and one exceeded its lead limit. Quantified residual solvents were below the study’s comparison limits; one detected solvent, hexanal, could not be quantified. [2]

The mixed results are essential to understanding the finding. The study did not test BPC-157, and it does not show that every online peptide is “95% impurities.” Method-specific purity measurements are not percentages of toxic material. Even a high purity result cannot resolve all questions about identity, metals, solvents, microbial contamination, or clinical safety. [2]

The semaglutide study: too much active drug and poor purity can coexist

A 2024 JAMA Network Open research letter examined purchases from six online sellers classified as rogue or not recommended by pharmacy-monitoring organizations. Only three purchases arrived; the others became nondelivery scams. Researchers tested the received products without administering them to patients. [3]

All three contained semaglutide. Reported purity was 7%–14%, against an advertised 99%, while the amounts of semaglutide were 29%–39% above the labels. One sample had elevated endotoxin, a bacterial toxin, even though no viable microorganisms were detected. [3]

Those findings are compatible: active-drug content compares the amount with the label, while purity describes a different analytical measurement. A preparation can contain too much active drug and still have poor measured purity. Likewise, failing to detect living microbes does not establish the absence of bacterial toxins.

The study excluded higher-priced offerings, and only three products were received. It cannot estimate a market-wide defect rate. It does show why finding the named drug in a sample is insufficient reassurance about the product. [3]

A counterfeit can contain a different drug

A case report first published online in October 2025 described a 31-year-old woman admitted with a coma caused by very low blood glucose after using a website-obtained product presented as semaglutide. Toxicological analysis identified insulin instead. The authors reported the incident to the Italian Medicines Agency and local authorities. [6]

This was one patient and a falsified product. It does not estimate how often such events occur, and insulin poisoning from a counterfeit should not be attributed to authentic semaglutide medicines.

A 2026 analysis of EudraVigilance safety reports also found different reporting patterns for suspected counterfeit semaglutide. Reports were grouped using codes for suspected counterfeiting or tampering, without systematic chemical confirmation of the products. Reporting bias, uncertain composition, and the unknown number of exposed users limit the conclusions. These reports can flag problems worth investigating, but cannot establish causation or a user’s probability of harm. [7]

Why one favorable test cannot establish safety

Professional laboratories can answer defined questions about submitted samples. That is useful for manufacturing, research, and investigations. It is a narrower task than establishing that an unfamiliar preparation is appropriate for someone to inject.

FDA’s March 2026 guidance on endotoxin testing discusses validated methods, sampling plans, manufacturing controls, and sample handling. These controls explain why a home check, photograph, or isolated favorable test cannot establish a product’s complete quality. A result depends on what was tested, how it was tested, and whether the sample represents the preparation someone would use. The guidance is nonbinding except where it cites applicable law or regulation. [8]

Endotoxins are components of certain bacteria that can remain after the bacteria have died. Sterility and endotoxin control therefore answer different questions. Trying to make an uncertain preparation usable at home does not resolve its chemical identity or clinical risks. [3] [9]

BPC-157 adds a problem of identity. In its May 11, 2026 scientific briefing for the July 23–24 compounding advisory meeting, FDA evaluated two chemical forms, the free base and acetate, and described inconsistent use of the name for salts and derivatives. It identified missing information about impurities, aggregates, and microbiological quality. Aggregates are molecules clumping together; this can affect a peptide preparation’s behavior and potential immune reactions. [5]

That briefing was a scientific evaluation, not a finding that all marketed BPC-157 products have the same defects or a final drug approval. Its practical lesson is that a familiar name may still leave the actual preparation inadequately specified. [5]

Compounding is a separate category

Research-labeled products, counterfeit branded medicines, and compounded preparations are different categories. Counterfeits misrepresent what they are. Compounding—the preparation of a medicine for a particular need—can serve a legitimate purpose, such as avoiding an ingredient a patient cannot tolerate. [10]

In the United States, compounded drugs are not FDA approved and do not undergo the agency’s premarket review for safety, effectiveness, and quality. Applicable conditions and oversight vary with the compounding setting. “Prescribed” or “compounded” does not itself establish FDA approval of that preparation. These are US distinctions; they do not determine Czech or EU legal status. [10]

How Healthy Longevity Clinic experts evaluate the evidence

For someone hoping to recover from an injury, manage weight, or improve health, the useful question is whether a defined treatment can help with that goal. Healthy Longevity Clinic’s interpretation separates two decisions: whether the preparation has adequate identity and quality controls, and whether clinical evidence supports its intended use. Passing one of those tests cannot supply the answer to the other. [2] [3] [5]

The semaglutide findings make this distinction concrete. Detecting the expected molecule did not resolve excess content or endotoxin. Conversely, a perfectly manufactured research compound would still need evidence of benefit and harm in people with the relevant condition. The clinical conversation should therefore start with the exact preparation, intended outcome, and available established options. [3] [8] [10]

An assessment would change with reliable quality documentation for that specific preparation and appropriate human studies showing meaningful benefit with acceptable harms. Another purity percentage, a laboratory result for a different sample, or a study of a differently formulated molecule would leave important questions open. [2] [3] [5] [8]

Three questions for a clinician

1. What exact medicine or investigational preparation is being considered, and what evidence addresses my condition?

2. What quality controls and regulatory oversight apply to this preparation, beyond its label or a single test result?

3. What established options could meet the same need, and how do their likely benefits and harms compare?

Common questions

Does a high purity number mean an injectable is safe?

No. It is a result from a particular analytical method. It does not by itself establish the correct amount, sterility, endotoxin control, or clinical benefit. [2] [3] [8]

Can a laboratory test tell me everything I need to know?

A laboratory can answer specific questions about a submitted sample. A single result cannot show that all relevant hazards were assessed, that another vial has the same contents, or that the treatment works for the intended condition. [5] [8]

What if I have already used a suspect product?

Tell a health professional what you took and retain the packaging for assessment. After a suspected falsified diabetes or weight-loss product, sweating, dizziness, or blurred vision warrant urgent medical attention; seizures or loss of consciousness require emergency help. The UK medicines regulator highlights these signs because falsified pens may contain insulin. In the US, suspected medication problems can also be reported through FDA MedWatch. [4] [11]