Are Research Peptides Making People Sick? What the 2026 Evidence Shows

What the headlines are based on

Most of the 2026 coverage traces back to three sets of facts: two adverse-event reports filed with the FDA, US poison-centre data, and a cluster of liver injuries in Australia. Each is real. Each also comes with limits that the headlines tend to leave out, and the gap between the two is where most of the confusion sits.

It helps to separate three questions. Did something bad happen to people using these products? Is there evidence the product caused it? And is there evidence the products are not what their labels say? The answers are, respectively, yes; not yet, in the cases reported; and yes, clearly.

The two US death reports

In October 2026, CBS News obtained two reports from the FDA's adverse-event reporting system through a records request.^[1]

• July 2026. A healthcare professional reported that a 41-year-old woman with obesity, otherwise described as healthy, went into cardiac arrest after taking retatrutide obtained "from an unreliable source". She was taken to hospital with a dangerously fast heart rate and died after resuscitation failed. The report also notes she had been drinking kava.
• September 2026. A 49-year-old woman died after severe dehydration, malnutrition and cardiac arrest. According to the report, a licensed nurse persuaded her to buy retatrutide from an online peptide seller and gave her dosing recommendations.

Neither report has been verified. Anyone can file an adverse-event report. Released reports are redacted, and a report cannot by itself show that a product caused an outcome. The FDA declined to say whether it had checked either one. CBS was clear that it is uncertain whether retatrutide, or something else in the vials, was to blame.^[1][2]

What can be said is narrower. Retatrutide is not approved anywhere, so no legitimate retail product exists. Anything sold under that name outside a clinical trial is unregulated by definition. Eli Lilly, which developed the drug, told CBS it has found products sold under the name with incorrect amounts, wrong ingredients and contaminants.^[1] And the second report describes the kind of harm that excessive GLP-1-type dosing is known to cause, which matters for the testing results below.

The poison-centre numbers

America's Poison Centers logged 249 exposures involving retatrutide in the first three months of 2026, 460 in the second quarter and 862 in the three months to September.^[1] An exposure is a call about someone who took or was exposed to a substance. It is not a confirmed poisoning, and many calls involve mild symptoms or simple questions.

Rising counts are also expected when a substance has only recently started being tracked, as awareness of the category spreads among callers and staff. The trend still points in one direction: more people are using these products, and more of them are running into problems serious enough to call for help.

The Australian liver injuries

The best-documented harm cluster is in Victoria, Australia. By June 2026 the state health department had confirmed six cases of acute liver toxicity in people who had used an unapproved product labelled as retatrutide, bought online, through friends or via social media advertising.^[3] By late August the count had reached at least seven, and some patients were described as seriously unwell with liver injury or liver failure. No deaths or transplants had been reported.^[4]

The cause has not been established. Victoria's Chief Health Officer said the clinical picture suggested that additional contaminants may be contributing, and a liver transplant specialist treating patients said the cause could not be confirmed until product testing came back. Liver injury of this kind is not a recognised effect of retatrutide in its clinical trials, which is one reason contamination is being considered. Australia's drug regulator, the TGA, has made unapproved peptides a compliance priority and reported destroying more than 94,500 unapproved peptide products in 2026.^[3][5]

What testing has found inside the vials

Harm reports raise the question. Testing answers part of it, and the testing data is the most solid evidence available.

Retatrutide vials in Australia

At the request of ABC News and the University of Queensland, an Australian testing laboratory analysed vials of retatrutide obtained from people using black-market products. Of three vials tested for the story, one contained less than labelled and two contained much more. Across the 18 retatrutide samples the same laboratory had tested in 2026, only one matched its label. About 30% were over-strength, about 30% were under-strength, and one contained no retatrutide at all.^[5]

The researchers were surprised by the direction of the error. With black-market steroids, under-dosing is the usual finding. With these peptides, over-strength vials were just as common. One of the researchers, writing in Drug and Alcohol Review, concluded that inaccurate dosing alone may be an important source of risk.^[5]

Thousands of third-party test results

A preprint posted in April 2026, not yet peer-reviewed, analysed a public database of third-party test results for 14 research peptides, including BPC-157, TB-500, CJC-1295, semaglutide, tirzepatide and retatrutide. After excluding incomplete reports, 6,285 results were analysed from more than 200 manufacturers.^[6]

• Typical results looked good. The median sample contained 101.8% of its labelled amount at 99.8% purity.
• The spread was the problem. Judged against a compounding-pharmacy standard (90-110% of label, at least 98% purity), 41.6% of samples failed. Against the tighter standard used for manufactured drugs, 71.1% failed.
• Some vials held the wrong thing. The labelled peptide was absent in 10.0% of TB-500 samples and 9.1% of CJC-1295 samples.
• Purity did not predict contamination. Only 243 samples were tested for endotoxin, a bacterial toxin that can cause fever, shock and, at high levels, death when injected. Of those, 15% had measurable endotoxin, and purity showed no relationship with endotoxin at all.

The authors note an important bias. These samples were sent in voluntarily, often by sellers who expect them to pass. The real failure rate in the market is unlikely to be lower than this, and could be higher.

Why a dose error is dangerous with these drugs

For many research peptides the main risk of a wrong dose is that nothing happens. For GLP-1-type drugs such as retatrutide, semaglutide and tirzepatide, the opposite is true. Their dose-limiting effects are gastrointestinal: nausea, vomiting and diarrhoea, which in clinical trials are managed by starting low and raising the dose gradually over weeks or months.

A vial that is 30% or more over-strength defeats that process. Each injection delivers a larger step than intended, and because these drugs are long-acting, the effect of a large dose persists for days. Persistent vomiting leads to dehydration and electrolyte disturbances, and those can affect heart rhythm. Reported cases in 2026 include a man hospitalised after a week of unceasing diarrhoea following an apparent overdose of an experimental GLP-1, and an Australian patient who tore their oesophagus from repeated vomiting.^[2] Neither case had confirmed product testing. Both are consistent with what over-dosing these drugs is expected to do.

See our week-by-week guide to GLP-1 side effects for how these effects present with approved drugs, and our retatrutide page for what its trials showed at controlled doses.

What a certificate of analysis can and cannot tell you

Sellers of research peptides commonly publish a certificate of analysis, and some buyers send samples for their own testing. Both are useful. Both have hard limits that the 2026 evidence makes plain.

• It describes one sample. A certificate reflects the vial that was tested, often a different vial, and sometimes a different batch, from the one in hand.
• Purity is not safety. The preprint above found no link between a high purity result and the absence of endotoxin. A vial can be 99.8% pure and still contaminated.
• Most certificates skip the tests that matter most for injection. Sterility, endotoxin and heavy metals are separate tests, and they are often missing.
• Identity needs its own test. Purity measures how much of the sample is one compound. Only an identity test such as mass spectrometry confirms it is the right compound.
• Nothing on paper replaces manufacturing controls. Approved and properly compounded drugs are made under rules that cover sterile processing, batch release and recalls. A test result is a snapshot, not a system.

Our guides to reading a certificate of analysis and judging a source go into each of these in detail. The short version is that documentation can screen out the worst products. It cannot certify that an unapproved injectable is safe.

Licensed supply chains are not immune

Two 2026 episodes show that the problem is ingredient quality and manufacturing control, not the label alone.

• Compounded glutathione. In August 2026 the FDA said it was aware of at least 30 patients with reactions consistent with excessive endotoxin after receiving intravenous glutathione compounded from dietary-supplement-grade material, and reminded compounders not to use that grade for injectables. Pharmacies recalled affected products.^[7] See our glutathione page.
• Counterfeit Ozempic. In December 2025 the FDA warned that counterfeit Ozempic pens had entered the legitimate US supply chain, after earlier counterfeit lots were found with fake components and non-sterile needles.^[8]

Regulation does not guarantee safety, but it does provide the means to find problems, trace them to a batch and pull the batch. That machinery is what the research-peptide market lacks. When something goes wrong with an unregulated vial, nobody is obliged to investigate, and the next buyer of the same batch never hears about it.

How regulators are responding

In the US, the FDA sent warning letters on August 24, 2026 to several online peptide sellers, rejecting "research use only" disclaimers on products marketed for weight loss and other effects.^[9] Eli Lilly has sued sellers of products marketed as retatrutide in four federal courts, as our coverage of the lawsuits explains. In Australia, the TGA and police have run seizures and the TGA warns against injecting any unapproved peptide.^[5]

Enforcement can shrink a market, but it does not change what is already in circulation. For anyone who has used one of these products, or is close to someone who has, the practical question is what to do if something goes wrong.

If you or someone else has a reaction

This section applies to anyone who has used an unapproved or research-grade peptide, whatever the source. It is general information, not a substitute for medical advice.

Get medical help promptly for

• Vomiting or diarrhoea that will not stop, or being unable to keep fluids down
• Signs of dehydration: dizziness on standing, very little or very dark urine, confusion
• Severe or persistent abdominal pain, which can signal pancreatitis or gallbladder problems
• Yellowing of the skin or eyes, dark urine with pale stools, or itching with fatigue, which can signal liver injury
• Fever, chills or shaking within hours of an injection, which can signal endotoxin or infection
• Spreading redness, heat or swelling at an injection site
• Hives, facial swelling, wheezing or difficulty breathing, which need emergency care
• A racing or irregular heartbeat, chest pain, or fainting

In the US, call 911 for an emergency. For advice, the Poison Help line at 1-800-222-1222 is free, confidential and open around the clock.^[11]

Tell the clinician everything

Say exactly what was used, what the label claimed, how much was taken and when, and where it came from. Clinicians treating these cases have asked for exactly this.^[3] People sometimes hold back because the product was unapproved. A clinician cannot treat what they do not know about, and the information matters more than how it was obtained.

Keep the evidence

Keep the vial, any remaining contents, the packaging and labels, and a record of where and when it was bought. Photograph them if you cannot keep them. Testing a product after a reaction is how contaminants are identified, as the Victorian investigation shows.

Report it

In the US, report the reaction to the FDA's MedWatch program, online or at 1-800-FDA-1088. Reports can be made by patients, family members or clinicians.^[10] In Australia, report to the TGA. Individual reports are how the patterns in this article came to light.

Lower-risk routes that now exist

Much of the demand for research-grade GLP-1 drugs came from price. That gap has narrowed. Manufacturer self-pay programmes in 2026 put approved semaglutide and tirzepatide products at roughly $149 to $449 a month, and Medicare beneficiaries may qualify for the $50-a-month GLP-1 Bridge. Our guide to what GLP-1 therapy costs in 2026 covers the options. For other peptides, the route with oversight is a licensed clinician working with a licensed pharmacy, where the compound can legally be prepared. Our explainer on approved, compounded and research-grade products sets out the differences.

None of these routes is risk-free. Approved drugs have side effects, and compounded products have had quality failures of their own. The difference is that someone is accountable for what is in the vial.

Our read of the evidence

The deaths in the headlines are unproven, and the reporting around them often outruns the documents. That should not be mistaken for reassurance. The testing evidence is consistent across two countries and thousands of samples: research-grade peptides often do not contain what their labels say, at the strength their labels say, and purity results do not rule out contamination. For GLP-1-type drugs, an over-strength vial can cause exactly the kind of harm described in the reports. The most useful thing anyone can take from 2026 is that no amount of paperwork makes an unregulated injectable predictable, and that anyone who has a reaction should seek care, keep the product, and report it.