Longevity Peptides: The FDA Panel Voted, and Not One Vote Was About Aging

The 300 percent problem

In January 2026, Americans ran 10.1 million Google searches about peptides. About 60% of those were people looking up GLP-1 drugs, which is fair enough — those are peptides, they are approved, and they very obviously do something. The rest is where it gets interesting: searches for so-called longevity peptides are up roughly 300% year on year, the peptide hashtag sits on more than 270,000 TikTok videos and 654,000 Instagram posts, and the claims attached to them include repairing DNA, regrowing hair, healing nerves and lowering stress.^[2]

CBS News chief medical correspondent Dr Jon LaPook summarised the evidence base for that last category with admirable economy: there is animal data suggesting powerful effects on cell function, “but there's no gold-standard reproducible randomized trials in humans that show they actually work.”^[2] Dr Monica Wang of the Harvard T.H. Chan School of Public Health added the line we keep taped above the desk: when something is marketed for everything, that is usually a red flag.

Then, in July, the FDA convened an advisory committee, and the peptide corner of the internet got the headline it had been waiting two years for.

What actually happened on 23 July

The Pharmacy Compounding Advisory Committee — PCAC, for people who enjoy acronyms — met over two days to consider seven peptides for the Section 503A Bulks List.^[1] That list matters in a narrow, technical way: it governs which raw substances a compounding pharmacy is allowed to make medicines from. It is not an approval pathway. Nothing on that list has been shown to work.

Six of the seven got a favourable vote. Here is the full slate, and we would encourage you to read the third column rather than the second.

Ulcerative colitis. Wound healing. Insomnia. Obesity. Stroke. Every one of these peptides is sold online as an anti-aging compound, and not a single one was assessed as one. Epitalon in particular is marketed almost exclusively on telomeres and lifespan; it was put in front of the committee for trouble sleeping.

This is not a technicality. It is the difference between “a panel looked at whether this could be compounded for a specific bowel disease” and “a panel endorsed this for slowing down aging.” Only one of those things happened, and it is not the one in your feed.

What the votes did not mean

Three further things got lost in translation. First, the committee is advisory: the FDA is not required to follow it, and actually adding these substances to the list requires notice-and-comment rulemaking that could stretch into 2027 or run for years.^[1] Second, the vote concerns whether pharmacies may compound with a substance, not whether the substance is effective. Third — and this is the part we find genuinely striking — FDA's own scientists told the committee there was “insufficient safety, efficacy, and moiety characterization information” for these peptides. The committee voted yes anyway, mostly by margins of two votes.

“Moiety characterization” is bureaucrat for we are not fully certain what molecule is in the vial. Hold that thought.

BPC-157: the one everybody actually means

BPC-157 (C₆₂H₉₈N₁₆O₂₂) is a synthetic 15-amino-acid sequence derived from a protein found in gastric juice. In rodents it does look remarkable — accelerated tendon and muscle healing, gut protection, effects on nitric oxide signalling, the lot. There are around 186 papers on it in PubMed.^[4]

So we ran the obvious search. PubMed, BPC-157 combined with the human and randomised-controlled-trial filters, on 10 August 2026. Result: zero papers.^[4]

Not “a few small ones”. Not “mixed”. Zero. A molecule with two decades of animal literature, a thriving grey market and its own subreddit has never had a published randomised controlled trial in humans indexed on PubMed.

ClinicalTrials.gov tells the same story from a different angle. Three records mention BPC-157:^[5][6]

• A Phase 1 safety and pharmacokinetics study in 42 healthy volunteers, run in Croatia between October 2015 and February 2016. Its registry status is still listed as unknown, ten years on.^[5]
• A 40-person study of a commercial peptide gummy, which is not the same thing as studying the peptide.
• A Phase 2 randomised, double-blind, placebo-controlled trial in 120 people with MRI-confirmed grade II hamstring strain. It started on 2 February 2026. Primary completion is due 14 February 2027.^[6]

That last one is genuinely good news and we will be reading it the day it lands. It is also the point: the first properly designed human efficacy trial of the internet's favourite healing peptide is currently still recruiting. Everyone injecting it today is ahead of the evidence by at least a year, and that is assuming the trial is positive.

TB-500 is not the molecule that was tested

This one deserves its own section because the sleight of hand is so easy to miss. TB-500 is a synthetic fragment of thymosin β4, a naturally occurring 43-amino-acid protein. When peptide sellers cite “human clinical trials” for TB-500, the trials they mean are almost always trials of full-length thymosin β4 — a different molecule.

And those trials exist. They are just not about what you would hope. The largest is a Phase 3 trial in 601 people — for dry eye disease, using thymosin β4 as an eye drop.^[7] The rest of the registry is a Phase 2 in 72 people with pressure ulcers, a Phase 2 in epidermolysis bullosa that was terminated, and a small run of cardiology studies in acute myocardial infarction.^[8]

So: different molecule, different route of administration, different problem. Citing an eye-drop programme as evidence for injecting a fragment to heal your shoulder is a bit like citing the crash-test rating of a Volvo when you are riding a scooter. Related industry, yes. Same protection, no.

The rest of the slate, counted honestly

We ran the same exercise across the others. These are our own literature counts, done on 10 August 2026, and you can rerun any of them in about fifteen seconds.

Notice the pattern. In every case the biology is plausible, the animal work is real, and the human efficacy evidence is somewhere between thin and absent. That is not the same as “these don't work” — it is the honest statement that nobody knows, including the people selling them. It is the same shape of problem we found with senolytics: spectacular in mice, still waiting in humans.

The risk nobody is discussing

Most supplement debates are about whether you are wasting money. This one is not, and the difference is the needle.

Peptides sold through research-chemical suppliers are not made to pharmaceutical standards, and are routinely shipped with a “not for human consumption” label that exists purely as legal cover. The concerns FDA reviewers raised in July were not vague: immunogenicity (your immune system reacting to a foreign protein sequence), peptide-related impurities, and incomplete characterisation of the active ingredient.^[1] Those are exactly the failure modes you cannot detect at home, and they land in your bloodstream rather than your stomach.

Put bluntly: with a dodgy capsule, the worst case is usually that nothing happens. With a dodgy injectable, the worst case is a sterile-technique problem or an immune reaction to a contaminant nobody characterised. The label-reading rules we use for supplements — third-party testing, certificates of analysis, identity verification — are not optional here; they are the entire safety margin, and the grey market largely does not provide them.

So what would change our mind

Something specific, and it is on the calendar. If the Phase 2 hamstring trial reports in 2027 with a clean positive result on MRI-confirmed healing, BPC-157 becomes a real drug candidate with real evidence, and we will say so loudly.^[6] Five more peptides are also queued for PCAC review before February 2027.^[1] This field could look quite different in eighteen months.

What will not change is the arithmetic in the meantime. There is a version of the peptide story that is completely true: GLP-1 receptor agonists are peptides, they went through the full trial process, and the evidence that they do something profound to metabolism — and possibly to biological aging — is now substantial. That is what a peptide looks like when it has been tested properly. It is also why 60% of those 10.1 million searches were about them.

The longevity peptides are the other 40%: the same chemical class, borrowing the credibility, without the trials. If you are building a routine on evidence rather than vibes, they belong in the same drawer as most of what we cover in the hype detox — interesting, unproven, and several years early. The short list of things that actually work remains short, boring and, for now, peptide-free. And the levers that are genuinely known to move healthspan are still free.