The GLP-1 Compounding Crackdown: What the FDA Has Done, What the Safety Data Says, and What It Means for Prescribers and Pharmacies

For two years, compounded semaglutide and tirzepatide were the largest compounding market the United States has ever had. At the 2024 peak, nearly one in three Americans taking a GLP-1 was getting it from a compounding pharmacy. That market was built on a shortage, and the shortage is over. What has followed in 2026 is the most concentrated enforcement campaign the compounding sector has seen: waves of warning letters, an FDA safety notice built on nearly 1,700 adverse event reports, and a proposal to shut the largest remaining legal pathway for good. This article lays out what actually happened, what the safety data does and does not show, and what it means for a prescriber or a pharmacy that wants to stay on the right side of the line.

This article is educational and is not medical or legal advice. It describes FDA actions and published evidence as of September 2026. Prescribing decisions belong to the clinician and the patient; compliance questions belong with your regulatory counsel.

How we got here: the shortage that made a market

Federal law gives compounded drugs a narrow lane. A state-licensed 503A pharmacy may compound a patient-specific prescription, but it may not regularly produce what is “essentially a copy” of a commercially available drug. A 503B outsourcing facility may compound in bulk without individual prescriptions, but only from substances on the FDA’s 503B bulks list or for drugs that are on the FDA’s shortage list at the time of compounding. When the FDA placed semaglutide and tirzepatide on the shortage list in 2022, both lanes opened wide, and telehealth companies built businesses on top of them.

The lanes closed on a schedule. The FDA declared the tirzepatide shortage resolved on December 19, 2024, and the semaglutide shortage resolved on February 21, 2025. Once a shortage ends, the 503B pathway for that drug ends with it, and a 503A pharmacy is back under the essentially-a-copy rule. A large part of the industry kept going anyway, and that is the gap the 2026 actions were designed to close.

What the FDA did in 2026

March: 30 warning letters to telehealth companies

On March 3, 2026 the FDA sent warning letters to 30 telehealth companies over how they marketed compounded GLP-1s. The two violations named most often were claims that implied the compounded product was the same as an FDA-approved drug, and advertising a compounded product under the telehealth company’s own brand without saying who actually made it. Commissioner Marty Makary called it “a new era” and said the agency was “paying close attention to misleading claims being made by telehealth and pharma companies across all media platforms.” It was the second batch of telehealth letters since the FDA began a broader campaign against misleading direct-to-consumer drug advertising in September 2025, a campaign that by the FDA’s own count had produced more letters in six months than in the prior decade.

April: the policy clarification, and more letters

On April 1 the FDA restated the rules for both kinds of compounder in plain terms. For 503A pharmacies: a compounded drug with the same active ingredient in the same or similar strength and route as a commercial product is essentially a copy unless the prescriber documents a significant difference for that patient, and the agency’s enforcement discretion extends only to a pharmacy that fills four or fewer such prescriptions in a calendar month. For 503B facilities: semaglutide and tirzepatide are on neither the bulks list nor the shortage list, so there is no lawful basis to compound them from bulk substance at all. A further round of warning letters that month went to companies marketing unapproved GLP-1, dual GLP-1/GIP, and triple-agonist products, and, per the FDA’s later safety notice, to active-ingredient distributors and outsourcing facilities as well.

April 30: the proposal to close the 503B door permanently

The largest action was the quietest. On April 30 the FDA proposed to exclude semaglutide, tirzepatide, and liraglutide from the 503B bulks list, published in the Federal Register on May 1. The bulks list is the mechanism by which an outsourcing facility can compound a drug that is not in shortage, and the test for getting on it is “clinical need.” After reviewing the nominations, the agency said it “did not identify sufficient evidence” of such a need, and it was explicit that backorders, convenience, and cost do not count. In Makary’s words: “When FDA-approved drugs are available, outsourcing facilities cannot lawfully compound using bulk drug substances unless there is a clear clinical need.” The comment period, originally to June 29, was extended to July 30. As of this writing the agency has not published a final determination. If it finalizes the exclusion, the only remaining 503B route for these three drugs would be a future shortage.

Two things the proposal does not do are worth stating, because they are often blurred. It does not touch 503A pharmacies, which were never governed by the bulks list. And it does not create a new prohibition for 503B facilities today; it converts the current situation, where the drugs are simply off both lists, into a permanent finding that they do not belong on the bulks list.

What the safety data says

The enforcement has a factual backbone, and it is worth reading carefully rather than in headline form. The FDA’s standing safety notice on unapproved GLP-1s, last updated September 1, 2026, reports that as of May 31 the agency had received 990 adverse event reports associated with compounded semaglutide and more than 730 associated with compounded tirzepatide, which is where the “nearly 1,700” figure in the trade press comes from. Some involved hospitalization. The agency is careful about what the number means: the reports do not establish that the product caused the event, and the total is almost certainly an undercount, because most state-licensed pharmacies are not required to report adverse events to the FDA at all.

What the reports do show clearly is where the risk concentrates. Stanford Medicine’s July review puts the number at 95 percent of the 2024 adverse events being related to dosing and administration rather than to the molecule. The mechanism is mundane. Approved semaglutide and tirzepatide ship in single-dose pens; compounded versions usually ship in multi-dose vials with a syringe, and the patient draws up the dose in units. Confusing units with milligrams, or milliliters with units, is easy, and the FDA has documented patients who “inadvertently administered doses at 10 to 20 times the prescribed amount.” Poison-control calls involving GLP-1s are up more than fifteen-fold since 2019. The agency also describes clinicians miscalculating doses, and prescriptions written outside the approved label: higher single doses, more frequent dosing, faster titration.

Beyond dosing, the clinical objections that Stanford’s endocrinologist Marina Basina, MD, and pharmacist Michael Blyumin, PharmD, raise are about the product itself:

Salt forms. Some compounders have used semaglutide sodium or semaglutide acetate, which are chemically different from the base semaglutide in the approved drugs and have not been shown safe or effective in people.
Additives. Vitamin B12 is the common one, marketed as reducing nausea. A 2026 study found the B12 can bond with the peptide and form a new molecule that exists in no approved product.
Untested dosage forms. Sublingual drops and dissolving tablets have no clinical data behind them. Blyumin’s line is blunt: “Nobody legitimate makes sublingual semaglutide.”
Degradation. Peptides break down with heat and handling. The FDA notes complaints about injectable products arriving warm or with inadequate cooling.

None of this means every compounded GLP-1 was dangerous. It means the failure modes are specific, predictable, and mostly fixable by a competent pharmacy and a careful prescriber, which is exactly the argument the FDA is making for why the shortcut is no longer justified.

What is still lawful, and what is not

The confusion in the market comes from treating “compounded GLP-1” as one thing. It is several, and the law treats them differently.

Bulk 503B compounding of semaglutide, tirzepatide, or liraglutide for non-shortage use has no legal basis today and, if the proposal is finalized, none in the future short of a new shortage.
A 503A pharmacy filling a patient-specific prescription that copies an approved product (same ingredient, same or similar strength, same route) is an essentially-a-copy problem unless the prescriber has documented a significant clinical difference for that patient. A pharmacy that does this regularly, beyond the four-a-month discretion, is exposed.
A 503A prescription with a documented, patient-specific reason, for example a patient who cannot tolerate an excipient in the commercial product or who needs a strength that is not manufactured, remains within the statute. The documentation is the difference between a lawful compound and a copy, and “the branded drug is expensive” is not a reason the FDA accepts.
Marketing is now its own exposure. Calling a compounded product “the same as” an approved drug, or selling it under a telehealth brand without naming the pharmacy, is what the March letters were about.

For a prescriber, the practical rule is that every compounded GLP-1 prescription now needs a reason that would survive being read aloud, written at the time, in the record. For a pharmacy, the practical rule is that volume is the signal regulators are watching: a pharmacy whose GLP-1 output looks like a manufacturer’s will be treated like one.

Where PEPTPlus and its pharmacies fit

Prescribers ask us where compounded GLP-1s stand on PEPTPlus, and the honest answer is that the platform was designed for the world the FDA is now enforcing, not for the one that just ended.

Eligibility review before a product is listed. A pharmacy cannot put a product in front of prescribers until it has passed review for the FDA’s compounding rules. Copies of commercially available drugs, unapproved salt forms, and untested dosage forms do not get through. Sublingual semaglutide is not in the catalog.
No volume incentive on either side. Practices pay a flat subscription and pharmacies pay a flat subscription. PEPTPlus takes nothing per prescription and nothing from what the patient pays for the drug, so nobody in the system is paid more when more scripts are written. That is the structural opposite of the telehealth model that drew the March letters.
The reason travels with the prescription. The prescribe workflow records the sig, the days supply, and the clinical rationale at the time of signing. When a compounded product differs from a commercial one, the difference is documented where a regulator would look for it.
Every fill is traceable. Orders route only to pharmacies licensed for the patient’s state, each fill carries its batch, and the certificate of analysis for that batch is attached to the order the patient can see. Dosing instructions and auxiliary labels are structured, not free text, which is the single most useful thing a platform can do about the dosing errors behind most of the 1,700 reports.
The pharmacy is named. Patients always see which licensed pharmacy compounded and dispensed their medication. There is no white-label layer between them.

The crackdown is not a verdict on compounding. It is a verdict on compounding as a substitute for manufacturing, sold as if it were the approved drug. The pharmacies that will still be here in two years are the ones that treated the shortage as a temporary exception rather than a business model, and the prescribers who will be comfortable under scrutiny are the ones whose records already explain themselves. If that describes your practice, you can start prescribing on PEPTPlus, or read how verified sourcing works first.