Retatrutide death reports: what FDA and poison data show

Two deaths have been reported to the FDA after use of products sold as retatrutide, and poison center exposure counts more than tripled in two quarters. Neither record proves retatrutide caused harm or says what was in the vials. The public lab record answers the vial question only in part.

Retatrutide death reports: what CBS News found

On 13 October 2026, CBS News investigative reporter Daniel Gilbert reported on two FAERS reports obtained under the Freedom of Information Act. The FDA redacted identifying details before release.

CBS calls these the first public US accounts alleging that unauthorized use of retatrutide, or products claiming to contain it, played a role in a death. Neither is connected to a clinical trial, according to the documents and Lilly. The first report also says the woman was drinking kava at the time of the emergency.

The second report carries a photo of a vial with a seller's name on it. A lawyer for the company's owner told CBS the vial does not match its current packaging, and that the redacted report holds no batch or lot number to check against its records, so it cannot show whether the company sold the product. The lawyer said the FDA had not contacted the company. No public record ties a lot to that report, so we do not name the company here.

The FDA told CBS it reviews adverse event reports and gathers more information when needed. It declined to say whether it had verified either report. Lilly said the reports "make clear" both people took black-market retatrutide, and that it has found suspect products with too much, too little or the wrong active ingredient.

What a FAERS report does and does not prove

FAERS was built to watch approved drugs after launch. Anyone can file: a doctor, a nurse, a patient, a relative, a manufacturer. The FDA's own FAERS dashboard notes set the limits.

A grey-market product adds one more gap. A report naming "retatrutide" records what the label or seller said, and nobody in either case is known to have tested the vial. A cardiac arrest after an untested product folds the product, the patient's health and anything else taken into one report.

FDA says a single report can trigger a closer look, and reports from unsupervised use are the only safety record that market produces. Pieter Cohen of Harvard Medical School told CBS the FDA should try to obtain and analyse the products both patients injected.

What a poison center "exposure" counts

CBS, citing America's Poison Centers, gave this series:

That is about 287 a month in the latest quarter, 3.5 times the first. CBS notes America's Poison Centers began tracking retatrutide late in 2025, and that a steady rise is common for a newly monitored drug. We could not find the series published by America's Poison Centers itself on 15 October 2026, so the figures above are CBS's.

America's Poison Centers defines an exposure case as actual or suspected contact with a substance. The National Poison Data System coding rules count contact "regardless of toxicity or clinical manifestation", including cases later found to be non-exposures and adverse reactions during normal use.

So 862 means 862 cases where someone reported contact with a product called retatrutide. It does not mean 862 poisonings, and it does not separate genuine retatrutide from a mislabelled vial. Our inference, not a finding: a series that rises this fast reflects more use and more awareness of the poison line together, and the counts alone cannot split the two.

What the lab record shows about the product in circulation

Neither death report says what was in the vial. Lilly's three failure modes (too much, too little, wrong ingredient) can be checked against public lab results, as of the dates below.

Identity. In our analysis of published certificates, 319 of 322 retatrutide results with an identity test confirmed retatrutide, across Kovera Labs, Janoshik and three vendor archives (tests 4 February 2025 to 16 July 2026). One vial held 0.00 mg. Vendors chose which results to publish, so that is a floor. Finnrick, which publishes results whether or not a vendor likes them, showed 31 identity failures in its 194 most recent classified retatrutide tests when we read its public page on 12 October 2026, about one in six.

Fill. The certificate record points to "too much" far more often than "too little". Of 307 results with a measured mass, the median vial held 110.2% of its label, 159 held more than 110% and 17 held 120% or more. Two held under 90%. Our purity and fill study finds the same overfill across most compounds.

On a vial with 120% of its labelled mass, the label understates the contents by a fifth. That is a point about label accuracy and traceability only. It is not guidance on use, and vialaudit gives none.

Purity. The median purity in the same pool was 99.86%. Purity does not measure contamination with endotoxin or microbes, which needs separate tests. Our endotoxin explainer covers why the gap matters for anything injected.

None of these results covers the vials in either death report. They describe products that sellers or Finnrick sent to a lab, not the vial a nurse sold or one from "an unreliable source".

What retatrutide's trials say

In Lilly's Phase 3 programme, the most common side effects were gastrointestinal. In TRIUMPH-2, published in The Lancet on 29 September 2026, 28% of participants on 12 mg had nausea against 8% on placebo. Our side-effects breakdown gives the dose table, and the TRIUMPH results page covers all four readouts. Lilly plans to file for approval in the first quarter of 2027. Trial participants were screened, dosed to protocol and monitored. Neither person in the death reports was in a trial.

Why the lot number matters

Neither the harm reports nor the lab record can link a reported injury to a lot. Lilly is suing sellers, with 11 suits filed since August, partly on the argument that bad reactions to bootleg vials will be blamed on its drug. A batch code that matches a public certificate is the only way a specific vial can be traced at all. Our COA verifier checks a report number against the Kovera and Janoshik records we mirror.

Adverse events after any product can be reported to the FDA through MedWatch. In the US, Poison Help is 1-800-222-1222.

Sources

• CBS News, Daniel Gilbert, "Two deaths were reported to the FDA. Is the bootleg weight-loss drug 'Reta' to blame?", 13 October 2026
• FDA, FAERS Public Dashboard, about and limitations
• FDA, Potential signals of serious risks identified from FAERS
• FDA, Adverse Event Reporting System overview
• America's Poison Centers, NPDS FAQs (exposure and information cases)
• National Poison Data System coding manual, case-type definitions
• Finnrick, retatrutide test results, read 12 October 2026
• FDA MedWatch