Evidence review

Peptide Therapy: What Is Approved, What Is Compounded, and What Is Not a Medicine

Three categories exist and are constantly conflated. A small number of peptides are FDA-approved drugs. A larger group is compounded by licensed pharmacies against prescr

The three categories

Approved peptide drugs. Tesamorelin, liraglutide, semaglutide and tirzepatide are peptides that went through full premarket review. Their evidence base is the trial programme attached to each.

Lawfully compounded peptides. Prepared by a state-licensed 503A pharmacy against a prescription for an identified patient, or by an FDA-registered 503B outsourcing facility. Not FDA-approved and not reviewed for safety or effectiveness, but inside the regulatory system with a licensed pharmacy and an identifiable prescriber.

Research chemicals. Sold with "research use only" or "not for human consumption" labelling. That labelling is a complete answer about regulatory status. No prescriber, no licensed pharmacy, no verifiable testing, no recourse.

The regulatory position changed on 23 July 2026

This section was materially wrong two days ago and is corrected here, because the direction of travel reversed between the briefing documents and the vote.

What FDA staff said. Agency reviewers recommended against adding all seven peptides under review. For TB-500 they reported being unable to identify a single human clinical study. On BPC-157 the briefing package cited a lack of evidence supporting effectiveness for ulcerative colitis, the indication under review.

What the committee did. Across 23–24 July 2026 the Pharmacy Compounding Advisory Committee recommended six of the seven substances under review — BPC-157, KPV and TB-500 at 8–6 with one abstention, MOTS-c at 7–5 with two, semax at 8–5 and epitalon at 7–5 with one. Only emideltide was rejected, at 6–7. It overrode its own agency's scientists on every substance it recommended. Reporters in the room described an audible reaction when the tally was read.

All eight of the committee's newly added temporary members voted in favour on the first three; six other members voted against and one abstained. The committee has faced scrutiny over members with conflicts of interest.

Two facts that change how to read this

These are indication-specific. Each substance was reviewed against one proposed use, not for blanket compounding. BPC-157 was reviewed for ulcerative colitis — not for tendon or joint recovery, which is what drives nearly all of its consumer demand. TB-500 was reviewed for wound healing. Compounding either for a sports injury would fall outside the reviewed indication even after a listing.

All seven came off Category 2 on 23 April 2026. That reclassification — removal from the list of substances that may not be compounded — is what made this review possible. The July votes were never a re-ban risk; the only question was whether a new lawful channel opens. Coverage framing this as peptides surviving a threat has the direction backwards.

What the vote does not do

Three distinct legal events are being treated as one across most coverage, and the distinction decides what is lawful today.

1. Removal from Category 2 — the list of substances FDA has flagged with significant safety concerns.
2. A PCAC recommendation — what happened on 23 July. Non-binding.
3. Placement on the Category 1 compoundable list — requires formal notice-and-comment rulemaking, which commonly takes 8 to 12 months.

Only the third makes compounding lawful. The FDA is not bound by the recommendation and has gone against this committee before. In 2023 the agency reviewed a batch of popular peptides and declined to add them, stating they may present significant safety risks — which is what pushed these compounds into the grey market they have occupied since.

So nothing changed legally on 23 July. What changed is the probability of a future change, and the market has spent two days reacting as though the two were the same thing.

Why the evidence question is harder here than for GLP-1s

Tirzepatide has SURMOUNT-1 through SURMOUNT-5, SURPASS-CVOT and SUMMIT — tens of thousands of randomised participants with published protocols and registry identifiers. Most peptides marketed for performance, recovery or longevity have nothing comparable.

What they usually have is mechanistic plausibility, animal data, small uncontrolled human series, and a large volume of testimonial. Those are not the same thing, and the gap between them is where most peptide marketing lives.

We apply the same rule here as everywhere else on this site: a claim resolves to a registry entry and a peer-reviewed publication, or it is labelled as unestablished. Animal data presented as proof of a human clinical effect is never treated as evidence.

What to establish before buying any peptide

1. Is this substance approved, lawfully compoundable, or neither?
2. Is there a named prescriber with a current state licence? Check the medical board, not the seller.
3. Is there a named pharmacy with a current licence? Check the pharmacy board.
4. Does any labelling say research use only? That resolves the question on its own.
5. What human trial supports the specific claim being made? Ask for the registry number.

Why this sits on a GLP-1 publication

Because the two markets overlap operationally. Providers offering compounded GLP-1s frequently list peptides in the same catalogue, and patients moving between them may not register that the regulatory status differs completely.

It is also information about the provider. A company selling both a lawfully compounded GLP-1 and an unapproved research peptide is operating on both sides of a line, and that is a fact about its compliance posture worth knowing before you trust its pricing or its clinical claims.

Questions readers actually ask