The case for reform: better information, not prohibition alone

The rule that came out of thalidomide is simple: the maker must show that a medicine works and is acceptably safe, and an independent body checks, before it is sold. Early results and confident experts are poor guides.

Laboratory results often do not hold up either: when scientists at one company tried to confirm 53 “landmark” cancer studies, they succeeded with 6.¹⁴ Even approved medicines can reach the market with uncertain benefit. Of 68 cancer uses approved by the EMA from 2009 to 2013, only 24 had shown longer survival at the time of approval; after a median of more than five years, 33 still had no clear evidence of longer or better life.¹⁵

We take these seriously, and our proposals are shaped by them.

• Information can look like approval. A careful page about an untested peptide may make it seem normal.
• Testing can become marketing. A seller could use a test result as a stamp of quality.
• Help can attract new users. Services for people who already use a product might draw in others.
• Money is limited. Public money for trials of unlicensed compounds is not spent on care known to work.
• A regulator’s strength is saying no. Kelsey kept thalidomide off the US market despite constant pressure from its maker.¹¹ Rules that bend to demand lose that strength.

Thalidomide did not disappear. It later proved useful in leprosy and in multiple myeloma, a blood cancer, and in 2008 it was authorised in the EU for myeloma.^10,12 But in rural areas without strong medical surveillance, its use by pregnant women with leprosy has kept causing birth defects.¹⁰

The same molecule can harm or help. Information, control and follow-up decide which. We think that lesson applies here too.

Seizures measure enforcement as well as trade, and side-effect reports only what someone chose to report. So these numbers show that use is real; they cannot say how large it is.

Most of the UK’s almost 20 million seized doses were not peptides: they were mainly sedatives, painkillers and erection medicines.¹⁶ Of 1,190 doping products seized at the Swiss border, 146 were labelled as peptide hormones or growth factors.³

• Semaglutide bought online without a prescription. A Hungarian university made test purchases, and three vials arrived. The powder was only about 8% to 14% semaglutide, although the labels claimed 99% purity. Yet each vial held about 29% to 39% more semaglutide than its label said.⁴
• Products sold as retatrutide, Australia. Three products held 51%, 165% and 190% of the amount on the label.¹⁸
• Falsified peptide medicines, Belgium. A public laboratory tested the ten falsified peptide medicines most often found on the Belgian market, bought from suspected illegal websites. It found amounts that varied widely, low purity, and arsenic and lead in some samples.¹⁹

Careful injecting cannot fix any of these. See product risk.

The US study interviewed 80 weightlifters, 43 of whom used anabolic steroids. The users rated doctors as no more reliable on steroids than friends, websites or their dealers, and 56% had never told a doctor about their use.⁷

The study is about steroids, not peptides, and is more than 20 years old; we found no comparable study of peptide users. We expect the same pattern, but nobody has measured it: when use is something to hide, people hide it from the person who could help most. In the Netherlands, doctors must keep what you tell them confidential.²⁰ See talking to your doctor.

A systematic review of 37 studies from 12 countries found that a median of 94% of side effects of medicines in normal care were not reported to the national reporting systems.⁵ For products used outside the medical system, the share is unknown. The FDA itself says side effects of compounded GLP-1 medicines are likely under-reported, because ordinary compounding pharmacies are not required to report them to it.⁶

Reports work when people make them. In EudraVigilance, reports about possibly counterfeit semaglutide mentioned low blood sugar more often than reports about the genuine medicine.¹⁷ That fits what regulators found when they tested fake pens: some contained insulin instead of semaglutide.²¹

The estimate comes mainly from smaller companies and varies with its assumptions.²² Nor is the incentive problem absolute. Under current EU law, once a medicine is authorised, other companies cannot rely on its trial data to sell copies for 10 years (in some cases 11), whether or not its maker holds a patent.²³ The EU agreed a reform of its medicines law in December 2025, with the new rules expected to apply from 2028.²⁴

So the barrier is cost and uncertain return. A compound with weak early evidence, many existing sellers and no clear patient group is unlikely to attract a sponsor, and that is often the kind people buy outside pharmacies. Without public money, “does it work, and what harm does it do?” may never be answered.

Wegovy has been authorised in the EU since 2022,²⁵ but in July 2024 the Zorginstituut advised against paying for it from Dutch basic insurance for obesity, citing unclear health gain, unclear long-term effect, no clear way of stopping, and costs of up to €1.3 billion a year.²⁶ Those are real concerns. A new assessment is under way: the Zorginstituut’s scientific advisory council first discussed it on 5 October 2026.²⁷ Meanwhile the CBG reports that these medicines are increasingly used as slimming aids without a prescription.²⁸ Whether the two are linked has not been measured.

Retatrutide shows the gap from the other side. In a phase 3, double-blind, placebo-controlled trial of 2,339 adults with obesity and without diabetes, funded by its maker, average weight loss at 80 weeks was 17.6% to 25.0%, depending on dose, against 3.9% with placebo Human trials(Evidence level: Shown in controlled human trials).²⁹ On 5 October 2026 it was not under EMA evaluation,³⁰ while products sold as retatrutide were already being made and seized.²

Once a substance is widely used, whether it works is a public health question. Most drugs tested in people never reach approval,¹³ which is why the test is worth paying for: a clear negative result protects people too.

In 2021 the EMA and national regulators started a pilot giving scientific advice to universities and non-profit groups studying new uses of older, off-patent medicines. The aim was evidence strong enough for a company to use in a later application.⁸ It covered only authorised medicines. We propose extending the idea, with public funding, to the few unauthorised compounds used most widely, chosen openly.

Since the 1990s the Dutch Ministry of Health has made drug testing possible for people who use drugs, through DIMS at the Trimbos Institute; by 2011 more than 100,000 samples had been handed in.³¹ A systematic review of 90 studies found that such services appear to influence what people intend to do and do, especially after an unexpected result Humans, limited(Evidence level: Seen in humans in case reports or small uncontrolled studies). Most studies were cross-sectional, so the evidence is limited, and they are about recreational drugs, not peptides.³⁷ Public laboratories already test peptide products that were seized or bought from illegal websites.^19,3

To stop testing becoming marketing, a peptide service should report problems, never approval: warnings about dangerous batches, nothing a seller could show as a certificate, and a plain statement that one test describes one sample, not the next vial.

In the Netherlands anyone can report a side effect to Lareb, also for products bought online, and Lareb does not pass the report on to your doctor.³² Lareb has said that people who use illegal medicines can come to it too.³⁸

The system exists; what is missing is trust and use. We propose saying clearly and often that such reports are welcome and confidential, and publishing what they show for unlicensed peptides. See if something goes wrong.

Dutch medicines law forbids stocking, offering, selling, supplying and importing a medicine without a licence, and advertising it; breaches can be fined heavily.^33,34,39 Using a product is not itself named as an offence, although importing one is, and the ban has no exception for own use.³³ See Dutch law.

We support those rules. Enforcement should concentrate on sellers, above all those who make false health claims. In the US, the FDA has warned companies that labelled products “for research purposes” but sold them for use in people, with dosing instructions.⁶ Users are better reached with information than with fear, because fear keeps them away from doctors and from reporting.

In 2014 the Endocrine Society, a medical society, noted that most users of performance-enhancing drugs are not athletes but people who lift weights, that their health risks are greatly underappreciated, and that observational studies are needed because trials cannot ethically copy such use.⁴⁰ In the Netherlands, a first outpatient clinic for people with health problems from anabolic steroids opened in 2010.³⁵

We propose practical guidance from professional bodies on which checks make sense for a patient who uses an unlicensed peptide, and why. A doctor who knows what to look for is someone a patient can tell. See monitoring.

A conditional marketing authorisation can be granted on less complete data for a medicine that meets an unmet need, when having it now outweighs the risk of what is unknown. It lasts a year at a time, and the holder must complete agreed studies.³⁶ But it assumes an applicant who will deliver full data later, usually a company. Dutch compassionate-use programmes can also only be started by a manufacturer.⁴¹

We propose routes a public body or university can use, with the same duties to keep collecting evidence, and enforced: EU cancer approvals show how long uncertainty can last after a medicine reaches the market.¹⁵ See legal routes to treatment.

On 23 July 2026 the FDA’s Pharmacy Compounding Advisory Committee (PCAC) voted 8–6, with one abstention, to recommend BPC-157 for the “503A list” of substances US pharmacies may use to compound medicines on prescription. It considered seven peptides and recommended six: BPC-157, KPV, TB-500, MOTS-c, Semax and epitalon. These results are as reported from the public meeting; the FDA has not posted minutes.^42,43 The vote is advisory and non-binding, and it went against FDA staff’s own proposal.^42,44 Legal compounding remains prohibited until the FDA formally adopts the recommendation and completes rulemaking.⁴⁵

None of these peptides is an approved medicine in the EU, the UK or the US; Semax is registered as a medicine in Russia.^9,46,47,48 A 503A listing would allow pharmacy compounding on prescription in the US. It is not a marketing authorisation, it says nothing about whether they work, and it changes nothing in the EU or the Netherlands.^49,50

The FDA told its committee that a product compounded from a listed substance is not FDA-approved and need not follow the usual manufacturing rules (only the ban on unclean conditions applies).⁴⁹ The pharmacy also need not report its side effects to the FDA.⁴⁹ That is access without the evidence and warning signals we argue for.