Compounded GLP-1s: what FDA safety reports show, and what they cannot

What the numbers are

The FDA page on unapproved GLP-1 drugs reports the adverse event reports it had received as of May 31, 2026: 990 for compounded semaglutide and more than 730 for compounded tirzepatide. The page presents them as reports submitted to the FDA and not as confirmed cases. It says some dosing-error reports led to hospitalization. In the sections the Atlas read, it does not say how many reports involved hospitalization.

The page was current as of October 1, 2026. The compounded versus brand-name page covers the rest of what the FDA says about these products. This page is about the reports themselves.

How to read a report count

The FDA explains how to read the counts from its adverse event system on its public dashboard page. In plain terms:

• A report does not prove the drug caused the event. The cause could be the underlying disease, another medicine or something else, and the FDA has not necessarily verified what the reporter saw.
• The same case can be reported more than once, for example by a consumer and by a manufacturer, and counts can change as data are updated.
• There is no denominator. The system does not record how many people used a product, so a count cannot say how common a problem is.
• Reporting is largely voluntary and many side effects are never reported. Press coverage and how long a product has been sold can change how many reports arrive.
• Counts cannot be used to estimate how often a side effect occurs or to compare the safety of different products.

Why compounded counts are probably low

The FDA page says federal law does not require state-licensed pharmacies that are not outsourcing facilities to report adverse events, so underreporting is likely. It also says it is not always possible to tell whether the drug caused an event or other factors contributed.

A 2025 study of a large primary-care dataset found documented compounded GLP-1 use in about 8% of patients while surveys suggested far more people use them. If clinicians often do not know a patient is on a compounded product, a problem is less likely to be reported as one.

What kinds of problems are reported

The FDA does not rank these problem types by number in the sections the Atlas read.

A timeline of FDA actions

Dates are as the FDA pages and news reports state them. A warning letter is an FDA allegation and not a court finding. Rules change, so check current sources.

What published studies found about compounded products

Several studies tested or described compounded and online products. Who did the work matters, so the table says.

Can the counts say compounded is riskier than brand-name?

No. The FDA says report counts cannot be used to compare the safety of different products, and it gives no count of how many people used each. The Atlas did not find a trial or controlled study that compared a compounded product with the approved one. What the counts and the published tests do show is specific: people have made measuring and calculation errors with vials and syringes, some products did not match their labels, and compounded products are not reviewed by the FDA before they are sold.

What the FDA advises

• Buy only from state-licensed pharmacies, and report problems to the FDA's MedWatch program.
• Patients should confirm with their prescriber or compounder how the intended dose is measured and given, and prescribers should be vigilant when setting doses and titration schedules.
• The compounded versus brand-name page lists the FDA's red flags, and the 503A pharmacy page explains how compounding works. The Atlas does not give mixing or measuring instructions.

Quick answers

How many adverse event reports has the FDA received for compounded semaglutide?

As of May 31, 2026, the FDA reported 990 for compounded semaglutide and more than 730 for compounded tirzepatide on its page about unapproved GLP-1 drugs. They are reports, not confirmed cases.

Does 990 reports mean compounded semaglutide is dangerous?

The count alone cannot say. The FDA says counts do not show how common a problem is, and they cannot be used to compare products. The reports do point to specific problems such as dosing errors.

Why does the FDA say the counts are likely too low?

Reporting is voluntary, and the FDA says federal law does not require state-licensed pharmacies that are not outsourcing facilities to report adverse events.

What problems does the FDA describe?

Dosing errors by patients and clinicians, doses above the label and digestive effects such as nausea, vomiting, diarrhea, abdominal pain and constipation. It says many reported events appear consistent with those seen with approved versions.

Have studies tested compounded GLP-1 products?

A few. Laboratory tests by the makers of the approved medicines found different impurity profiles, and a small test of illegal online sellers found products that did not match their labels. The Atlas found few independent tests of products from licensed pharmacies.

Has a trial compared compounded and approved products?

Not that the Atlas found.

Where do I report a problem?

The FDA directs problem reports to its MedWatch program.

What did the FDA's 30 warning letters in 2026 say?

Fierce Pharma reports they cited claims that compounded products had the same ingredients as the approved drugs, had been tested in trials or implied FDA approval. A warning letter is an allegation by the FDA and not a court ruling.

Questions to ask

• Is the product you are giving me FDA-approved or compounded, and who makes it?
• Does the strength on the label use the same units as the instructions I was given?
• Who will show me how the amount is measured, and who do I call if the instructions confuse me?
• How will we know if I have a problem, and how do I report it?
• What happens to my plan if the pharmacy or the rules change?

What is still unknown

• The counts come from voluntary reports. They cannot show how often a problem happens or whether the product caused it.
• The FDA's counts do not say how many people used compounded products, so no rate can be calculated.
• The Atlas did not find a trial or controlled study comparing a compounded product with the approved one.
• Two of the published product tests were done by makers of the approved medicines. The Atlas found few independent tests of products from licensed pharmacies.
• The test of online sellers involved illegal sites and a small sample. It does not describe licensed pharmacies.
• Counts, rules and court cases change. The counts here are as of May 31, 2026, and the FDA page they come from was current as of October 1, 2026.