Peptide Therapy: Mechanism, Evidence and What the 2026 FDA Panel Vote Actually Decided

In short: across 23 and 24 July 2026 the FDA's Pharmacy Compounding Advisory Committee reviewed seven substances nominated for the 503A Bulks List and recommended six of them — BPC-157, KPV and TB-500 each by eight votes to six with one abstention, MOTS-c by seven to five with two abstentions, along with semax and epitalon; emideltide was rejected. None of that is approval. The list governs what a United States compounding pharmacy may use, not whether a medicine is safe or effective, the recommendation is not binding, and compounding remains impermissible until notice-and-comment rulemaking concludes. The substantive objection came from FDA staff and it was about chemical identity: these peptides lack universally accepted definitions, which leaves any safety assessment without a subject. The published science says the same thing in its own vocabulary. A 2025 systematic review of BPC-157 in orthopaedic sports medicine found 36 studies of which 35 were preclinical, and a 2026 review in Pharmaceutics concluded that the barrier to clinical translation is not absent biological activity but absent pharmaceutical science. Meanwhile the 503A list has no legal force in the United Kingdom or the Gulf, and BPC-157 remains prohibited in sport at all times under WADA class S0.

This Journal has spent the series asking whether a route earns its claim. Peptides pose a harder question, and an earlier one. Before route, before dose, before evidence, there is identity — and for most of the molecules sold as peptide therapy, that is where the chain breaks.

What "peptide therapy" refers to

A peptide is a short chain of amino acids. That is a description of chemistry, not of a therapeutic class, and it covers substances with nothing else in common. Insulin is a peptide. So is semaglutide, so is teriparatide, so is the carnosine dipeptide found in skeletal muscle. Each of those has a defined structure, a characterised pharmacokinetic profile and, where relevant, a marketing authorisation.

The phrase "peptide therapy" as used commercially means something narrower: a group of synthetic sequences sold outside that framework, usually for injection, usually labelled for research use only. The headline names are BPC-157 and TB-500 for tissue repair, MOTS-c for metabolic signalling, epitalon for sleep and longevity, semax and selank for cognition and anxiety, and the growth hormone secretagogues CJC-1295 and ipamorelin.

"Research use only" is the load-bearing phrase. It is not a safety grade or a purity standard. It is a statement that the material was not manufactured for administration to humans, and it is what allows a supply chain to exist alongside the medicines system rather than inside it.

What the July 2026 panel voted on

The Pharmacy Compounding Advisory Committee met on 23 and 24 July 2026 to consider substances nominated for the 503A Bulks List. On the first day it took up BPC-157, nominated for ulcerative colitis; KPV, for wound healing and inflammatory conditions; TB-500, for wound healing; and MOTS-c, for obesity and osteoporosis. On the second day it considered emideltide, also called delta sleep-inducing peptide, nominated for opioid withdrawal, chronic insomnia and narcolepsy; semax, for cerebral ischaemia, migraine and trigeminal neuralgia; and epitalon, for insomnia.

The margins were narrow. BPC-157, KPV and TB-500 each passed by eight votes to six with one abstention. MOTS-c passed seven to five with two abstentions. Semax and epitalon were also recommended. Emideltide was not.

What that vote does is worth stating precisely, because most of the reporting around it did not. The 503A Bulks List determines which bulk drug substances a compounding pharmacy in the United States may use when preparing a preparation for an individual patient. Inclusion is not a marketing authorisation, carries no finding of efficacy, and follows a different legal test from drug approval. The committee's recommendations are non-binding; the Secretary of Health and Human Services must act on them, and the FDA must complete notice-and-comment rulemaking — commonly estimated at eight to twelve months — before pharmacies have clear authority to compound these substances at all.

One committee member raised the obvious downstream risk during the meeting: that the public would read a procedural vote as an endorsement of efficacy, against a background in which placebo response rates of roughly 30% are ordinary. That is precisely what followed.

The question that stopped the room

The agency's reviewers did not argue the peptides were dangerous. They argued something more fundamental — that the substances lack universally accepted chemical definitions, and that this absence undermines any assessment of safety. The FDA's Russell Wesdyk put the difficulty in one line: the agency had never before faced a problem of what is it.

That is not a bureaucratic complaint. A peptide name is not a specification. Two vials labelled with the same three characters can differ in sequence length, in salt and counter-ion, in N-terminal or C-terminal modification, in stereochemical purity, in degree of aggregation and in impurity profile — and every one of those variables moves both the pharmacology and the toxicology. Safety data generated on one material do not transfer to another material sharing only its trade name.

A committee member voting in favour described the same gap from the other side: he was voting on something, without knowing what that something was. It is unusual for a regulatory record to state the problem so plainly, and it is the single most useful thing to take from the meeting.

BPC-157: thirty years of animals, one clinical record

BPC-157 is a synthetic pentadecapeptide derived from a fragment of a protein found in gastric juice. It has been studied since the early 1990s and the preclinical literature is genuinely large and genuinely consistent — angiogenesis, tendon and ligament healing, gastrointestinal protection, effects reported across many organ systems.

The human literature is a different matter. A 2025 systematic review in HSS Journal examined BPC-157 in orthopaedic sports medicine and identified 36 studies. Thirty-five were preclinical. The single clinical item was a retrospective assessment of musculoskeletal pain after intra-articular injection for unspecified chronic knee pain. There was no randomised placebo-controlled trial available to include, for any indication.

The preclinical work spanned intramuscular, intraperitoneal, intravenous and oral administration at doses from 6 micrograms per kilogram to 20 milligrams per kilogram — a range spanning more than three orders of magnitude, which is itself a sign that no dose has been established. The review's recommendation to clinicians was not enthusiasm. It was to ask patients directly about supplement use and to counsel them on the risks of unregulated manufacturing, contamination and poor clinical safety data.

Consistency in animals is a reason to run a trial. It is not a substitute for one, and this is the same discipline the series applied to intravenous NAD+, where mechanistic plausibility also ran far ahead of the clinical record.

Stability is not bioavailability

A 2026 review in Pharmaceutics examined BPC-157 as a candidate drug rather than as a phenomenon, and its findings are the most useful technical material published on the molecule.

BPC-157 does show unusual stability in gastric juice, resisting enzymatic and acidic degradation in a way uncommon among peptides. That property is the origin of the claim that it works orally. The review draws the distinction the claim elides: gastric stability is necessary for oral bioavailability but not sufficient for it, and oral bioavailability has not been characterised in any species under standardised fed and fasted conditions.

The rest of the profile is similarly incomplete. Plasma half-life is under 30 minutes — roughly 15 minutes in rat and 5 minutes in dog by the intravenous route, with a human pilot also under 30 minutes. Intramuscular bioavailability ranges from 14% to 51% depending on species. Subcutaneous pharmacokinetics, volume of distribution and plasma protein binding are simply absent, and subcutaneous injection is how the substance is overwhelmingly used. Dosing regimens in circulation derive from animal studies without formal allometric scaling or pharmacokinetic validation.

The review's conclusion deserves to be the last word on the subject: the primary barrier to clinical translation is not the absence of biological activity, but the absence of fundamental pharmaceutical science — characterised formulations, validated pharmacokinetics and a coherent development strategy. That sentence explains both why the preclinical literature looks impressive and why it has produced nothing approvable in three decades.

TB-500, MOTS-c, KPV, semax, epitalon

TB-500 illustrates the identity problem better than any argument about it could. It is sold under the name of thymosin beta-4, a 43-residue protein. Analytical work in the doping-control literature characterised the active component of TB-500 by liquid chromatography and high-resolution mass spectrometry as the N-terminally acetylated 17 to 23 fragment of thymosin beta-4 — the seven-residue actin-binding sequence LKKTETQ, with an acetyl group added. Seven residues are not 43, and an acetylated fragment is not its parent protein. The preparation itself originated in veterinary use.

MOTS-c is a mitochondrial-derived peptide with a real and interesting biology, and the human literature is almost entirely observational — circulating MOTS-c studied as a biomarker in dialysis and metabolic cohorts rather than administered as an intervention. KPV is a melanocortin-derived tripeptide corresponding to the final three residues of alpha-MSH; it acts through the PepT1 transporter and reduces inflammation in murine colitis models, with a dimeric analogue under clinical investigation. Both were nominated for indications — obesity and osteoporosis, wound healing — considerably broader than their evidence.

Semax and epitalon carry an additional complication. Both were developed in Russia and have regulatory recognition there, which is often presented as approval without qualification. It is not equivalent to an FDA or EMA authorisation, and the trial evidence underlying it is not generally available to Western reviewers in a form that permits independent appraisal. GHK-Cu, frequently grouped with these, sits in a different category again: its meaningful human evidence is topical and cosmetic, which says nothing about injection.

What an approved peptide looks like

None of this is an argument that peptides cannot be medicines. The contrast makes the point sharper than a caution would.

Tesamorelin is a 44-amino-acid analogue of human growth hormone releasing factor, modified with a hexenoyl group at the N-terminus to extend its half-life. It was approved by the FDA in 2010 for the reduction of excess abdominal fat in patients with HIV-associated lipodystrophy. It has a defined structure, a manufacturing specification, an established dose, a known adverse event profile and a narrow licensed indication supported by controlled trials.

That is what the full pathway produces: not a promise that the molecule is powerful, but a precise statement of what it does, in whom, at what dose, with what risks. Thymosin alpha-1 occupies an intermediate position — authorised in a number of countries for specific infectious and immunological indications, not approved in the United States or the United Kingdom — which is a useful reminder that "approved" is always a question of where.

What is actually in the vial

The identity problem is not confined to nomenclature. It reaches the product.

The most rigorous published data come from an adjacent market. A 2024 market-surveillance study in the Journal of Medical Internet Research purchased peptide products from online sellers operating without prescription and subjected them to laboratory analysis. Vendors claimed at least 99% purity. Measured purity in the three vials was 7.7%, 8.97% and 14.37%. Peptide content exceeded the labelled amount by between roughly 29% and 39%, which for a self-administered injection is a dosing error in the dangerous direction.

The microbiological findings are the part clinicians should read twice. All three lyophilised samples were free of viable microorganisms at testing — but bacterial endotoxin was detected in all three, in one case at 8.95 endotoxin units per milligram. Sterility and endotoxin are different tests. A vial can be sterile and still deliver a pyrogenic load, because endotoxin is a heat-stable cell wall fragment that survives the death of the organism that produced it.

That study examined semaglutide products rather than BPC-157, and the distinction should be kept. What it establishes is the behaviour of the unregulated peptide supply chain, and there is no reason to expect a research-use-only vial of a substance with no pharmacopoeial monograph at all to be held to a higher standard than a counterfeit of an approved medicine.

The 503A list is a United States instrument

Readers in the United Kingdom and the Gulf are being served American regulatory news as though it described their own position. It does not.

The 503A Bulks List exists under United States federal law and governs United States compounding pharmacies. It confers nothing in Britain or the Emirates. The nearest United Kingdom mechanism is the supply of unlicensed medicinal products — specials — under regulation 167 of the Human Medicines Regulations 2012, which permits supply to meet the special needs of an individual patient where no licensed product does.

The architecture is different in a way that matters. There is no list of permitted substances to point to, and the MHRA is explicit that the quality, safety and efficacy of unlicensed medicines have not been assessed by the agency and fall under the direct responsibility of the prescribing doctor. Where the American debate is about what a pharmacy may stock, the British position places the burden squarely on the individual prescriber — and a favourable advisory vote in Maryland does not lighten it by a gram.

Anti-doping: a separate and stricter line

For any clinic seeing competitive athletes, there is a further check that the medicines framework does not cover.

BPC-157 is prohibited under WADA class S0, non-approved substances — a class covering any substance not approved by a governmental regulatory health authority for human therapeutic use, prohibited at all times, in and out of competition. USADA states that BPC-157 is not approved for human clinical use by the FDA or any other global health authority, and that because it has not been extensively studied in humans, it is unknown whether there is a safe dose or any way to use it safely.

Class S0 turns on approval status, not on evidence of performance enhancement, and it has no out-of-competition window. A compounding rule change in the United States would not alter an athlete's exposure, and growth hormone secretagogues sit in a prohibited class of their own.

Evidence graded by peptide and status

Sorted by molecule, the field contains one fully approved agent, one regionally approved agent, several substances with substantial animal data and no controlled human trials, and a supply chain with documented quality failures.

How to read a peptide claim

Three questions separate a characterised molecule from a name on a vial.

First, what exactly is the substance? Not the trade name — the sequence, the terminal modifications, the salt form, the specification it is manufactured against and the analytical method that confirms it. If that question cannot be answered from documentation, the FDA's reviewers have already explained why nothing downstream of it can be assessed either. TB-500 is the worked example: the name says thymosin beta-4 and the vial holds seven residues of it.

Second, what does the evidence actually cover, and by which route? Animal data across three orders of magnitude of dose do not define a human dose. Topical evidence does not license injection. A retrospective note on knee pain is not a trial. This is the same error the series identified in biotin's six-month nail studies, in collagen's twelve-week oral protocols, and in bedtime glycine trials cited for afternoon infusions.

Third, which jurisdiction is being described? An American advisory committee vote is not a British or Emirati legal position, and it is not an approval anywhere.

Peptides are not the problem. Some of the best-characterised molecules in the pharmacopoeia are peptides, and the To Infinity MD concept in the IVIXIR series contains one — carnosine, a dipeptide with a defined structure, a known enzymatic fate and a documented serum half-life. The difference between that and a research-use-only vial is not ambition. It is that the question what is it has an answer. Holding that line — identity before pharmacology, pharmacology before evidence, evidence before claim — is the standard EFBA applies. This category tests it earlier than any other, because here the failure comes before the science has a chance to begin.