Three GLP-1 safety stories broke within four weeks this fall. An FDA-registered drug compounder in North Carolina recalled tirzepatide vials because it couldn't assure they were sterile.1 FDA inspection reports described black mold and insects at a Chinese plant that makes ingredients for compounded GLP-1 drugs.2 And CBS News reported that FDA had received 2 death reports involving products sold as retatrutide, Eli Lilly's investigational triple agonist. No company can legally sell retatrutide in the United States, and neither report has been verified.3
The common reaction online is a shrug: approved drugs get recalled too, big drugmakers get written up too, so you might as well buy what you can afford.
That shrug misses what a recall is. The manufacturing failures are alarming. The ability to identify affected lots and warn patients is the reassuring part. A recall documents action on an identified problem, started by the company or at FDA's request; it doesn't prove every safeguard worked.4 But when a vial comes from a website that insists no human was meant to use it, there may be no lot record to trace and no list of patients to warn. So who recalls a research peptide?
| Clinical Takeaway | ||||||||||||
|
Disclosure: I'm chief medical officer of Vineyard, a telehealth obesity practice. I provide research consulting to Novo Nordisk, whose FDA warning letter is discussed below, and consult for Viking Therapeutics. Full disclosure at the end.
What surfaced this fall
Apothecary Pharma, an FDA-registered 503B outsourcing facility in Cary, North Carolina, began a voluntary recall of compounded tirzepatide 18 mg/mL multidose vials on September 24 for lack of assurance of sterility. FDA has since classified it as a Class II recall, the middle of its three risk tiers.1,4 The same facility had received an FDA warning letter in December 2025 citing serious deficiencies in its sterile drug production, along with adverse event reporting procedures the agency called inadequate.5 This recall came from the most closely regulated tier of compounding.
A day later, Houston Rx, a Texas compounding pharmacy, began recalling 64 vials of tirzepatide mixed with niacinamide for the same reason.6 A month earlier, FDA had inspected the pharmacy and listed 14 observations.7 Inspectors saw an operator with exposed skin on the forehead stoppering freshly filled vials inside an ISO 5 hood, the cleanest workspace in a sterile pharmacy. On August 3, a surface sample from the ISO 5 area grew a microbial colony after a tirzepatide-niacinamide lot was filled; the pharmacy didn't investigate, and the lot was released and partly distributed. That's an environmental finding, not a test showing a finished vial was contaminated. The observation that matters most is about the ingredients.
| From the FDA inspection report |
| Form FDA 483, Houston Rx LLC, Houston, Texas. Issued August 28, 2026. A 483 lists inspectors' observations; it isn't a final agency determination, and the firm can respond. |
| Observation 9. "Use of ingredients not intended for pharmaceutical use in sterile drug production. Specifically, your firm is using active pharmaceutical ingredients and excipients that are not always USP/NF grade or are non-pharmaceutical grade." |
| The observation lists certificates of analysis describing the pharmacy's tirzepatide and semaglutide ingredients as "R&D only." |
| Source: FDA, Form 483 for Houston Rx LLC (PDF).7 |
At this pharmacy, at least, ingredients labeled for research went into sterile drugs.
Upstream, it gets worse. In August, FDA inspectors at Chengdu Shengnuo Biopharm, a Chinese maker of ingredients for compounded GLP-1 drugs, documented black mold in pipes and ceilings, significant insect activity, and quality-control staff pouring sample solutions into a sink in what inspectors called a coordinated effort.2 At Nanjing Chengong Pharmaceutical, inspectors found workers had stamped their own company's approval on ingredients made by other factories.8 Both plants are now on FDA import alert.8–10
Then there are the death reports. In July, a healthcare professional reported that a 41-year-old woman with obesity went into cardiac arrest and died; the report lists a product identified as retatrutide, obtained "from an unreliable source," as a suspect product and notes kava use. In September, FDA received a report that a 49-year-old woman died after severe dehydration, malnutrition, and cardiac arrest. According to the report, a licensed nurse had sent her a link to an online peptide vendor, sold her vials, and recommended doses.3
Neither report has been verified, and nothing public shows that anyone tested what was in those vials. Lilly told CBS it has found suspect products "with too much, too little, or the wrong active ingredient," along with contaminants.3 When University of Queensland researchers and the harm-reduction group PED Test Australia had a "research use only" vial labeled 10 mg of retatrutide tested, it held 19 mg, ABC News reported in June.11