An NCT number is the most authoritative-looking object in peptide research. It is short, official, government-hosted, and it converts an unproven compound into something that sounds like it is being properly studied. Vendor pages cite them. So do we.
This week that broke. A registration cited in yesterday's BPC-157 vs TB-500 comparison does not survive inspection, and the batch it came from is instructive enough to be worth an entire roundup. What follows is a correction, the verification method that produced it, and the two 2026 registry entries that do hold up.
The correction
Yesterday's post described NCT07437547 — a Phase 2 trial of BPC 157 in MRI-confirmed grade II hamstring strain, listed as recruiting since February 2026 — as the first adequately designed musculoskeletal efficacy test of the compound. We hedged it ("a registration is not a result") but still treated the record as a bona fide ongoing trial.
That was wrong, and we withdraw it. The record belongs to a set of eight registrations under a single sponsor name, several of which openly declare themselves to be examples rather than trials. The BPC-157 record itself contains no such declaration, which is precisely why it read as genuine. Yesterday's post has not been altered; this is the correction of record.
Nothing else in that comparison depends on it. The mechanism sections, the half-life correction, and the point that TB-500's human literature is almost entirely about full-length thymosin β4 rather than the heptapeptide all stand independently. What changes is the conclusion: BPC-157 still has no adequately designed, verifiable, ongoing efficacy trial. The human file is thinner than we said, not thicker.
What the batch actually contains
All eight records were pulled directly from the ClinicalTrials.gov API rather than from a search-result summary. They share a fingerprint:
- One sponsor name, "Hudson Biotech," first posted between February 27 and March 19, 2026, all listed as recruiting
- One study location across all eight — Peking University Shenzhen Hospital
- One central contact, the same name, phone number and email address on every record
- All eight flagged as not FDA-regulated drug studies
Three announce their own status. The TB-500 record (NCT07487363) opens: "This fictional study is an example of a ClinicalTrials.gov-style record." The Melanotan II record (NCT07437560) begins "This example interventional study record describes…" The tesamorelin record (NCT07481734) carries "(Mock Study)" in its own public title.
Two more are copies of somebody else's work. NCT07481747 reproduces the title of SURMOUNT-1 — Eli Lilly's tirzepatide obesity trial, which is NCT04184622, enrolled 2,539 participants, and completed in July 2024 — and carries Lilly's internal study identifier, I8F-MC-GPHK, with a suffix appended. NCT07467447 is registered against LY3437943, Lilly's compound code for retatrutide, under a second Lilly-format identifier. Neither is a new study.
The remaining three — the BPC-157 hamstring record, a Phase 2a of MOTS-c in prediabetes (NCT07505745), and a topical GHK-Cu wound-healing record (NCT07437586) — contain no self-declaration at all. They are written in ordinary protocol language, with plausible endpoints, masking descriptions and completion dates. Read one on its own and there is nothing to catch.
That mix is the whole lesson. A batch that is partly self-labelled and partly not means the labels cannot be relied on to travel with the records. The three quiet ones are the ones that end up in citations.
A registration is a submission, not a finding
This is a structural point about the registry that is easy to forget because the site looks like a database of research rather than a database of claims about research.
Records on ClinicalTrials.gov are submitted by sponsors and investigators. The site's own disclaimer states that listing a study does not mean it has been evaluated by the U.S. federal government. There is basic quality-control review, but no verification that a described trial is running, that the sites named are enrolling, or that the sponsor holds any rights to the compound.
So an NCT number tells you that somebody filled in a form. It is the beginning of a paper trail, not evidence. On our evidence hierarchy, a registration sits below rung three, not on it — it is not a human result, it is an intention to produce one, and intentions are cheap.
What a real 2026 registration looks like
For contrast, from the same registry in the same year: NCT07586865, a multicenter randomized, double-blind, placebo-controlled Phase IIc study of recombinant human thymosin β4 (NL005) in acute myocardial infarction, sponsored by Beijing Northland Biotech, first posted May 14, 2026, estimated enrollment 189, with co-primary endpoints of absolute and relative myocardial infarct size at day 90 and Fuwai Hospital (Chinese Academy of Medical Sciences) among its sites.
What makes this checkable is not the record — it is everything around it. The same sponsor has completed Phase 1a and 1b studies (NCT04555824, NCT04555850) and two prior Phase 2 studies in acute MI (NCT05485818, NCT05984134), and a first-in-human Phase 1 from that program was published in J Cell Mol Med in 2021. There is a program, with a history, moving in a direction.
Note also what it is: full-length recombinant thymosin β4, not the Ac-LKKTETQ heptapeptide sold as TB-500. The registry continues to be about the parent protein.
The registry news that matters most this year
The single most consequential 2026 entry for anyone tracking the research-peptide shelf is NCT07531251, nicknamed 4TAZPower: a randomized, double-blind, placebo-controlled trial of elamipretide in genetically confirmed Barth syndrome, 72 weeks, estimated enrollment 48, actual start July 2, 2026, primary completion estimated September 2029. The sponsor announced first patient dosed on July 8, 2026 under its new name, Mighty Therapeutics — the company formerly known as Stealth BioTherapeutics, renamed in June 2026, though the registry record still carries the old name.
This matters because SS-31/elamipretide received FDA accelerated approval in September 2025 for Barth syndrome, on an intermediate endpoint (knee extensor muscle strength). Accelerated approval is conditional by design: continued approval depends on verifying clinical benefit in a confirmatory trial. 4TAZPower is that obligation being discharged.
The pattern is worth internalising. The first mitochondria-targeted drug ever approved is, four years from now, still scheduled to answer the question its approval was granted in anticipation of. That is what an unfinished evidence story looks like when it is being handled properly — a named trial, a defined endpoint, a completion date, and a regulator holding the marketing authorisation against it.
One genuine result, correctly placed
Not everything this year is registry mechanics. Bian et al., eBioMedicine 2026;124:106124 (published January 20, 2026; PMID 41564845) reported that the senolytic combination dasatinib plus quercetin, given as a short oral "hit-and-run" course in streptozotocin-diabetic mice, improved kidney function and reduced markers of injury, fibrosis, inflammation and p16-positive senescent-cell burden, while raising the geroprotective factors α-Klotho and Sirtuin-1 — without altering glucose.
Two placements. First, this is rung two: a mouse model of a construct, with supporting work in human cell lines. It is a well-executed animal study, and it is an animal study. Second, neither dasatinib nor quercetin is a peptide — this is the senolytic concept accumulating evidence, which is the concept FOXO4-DRI belongs to, not evidence about that peptide. Class-level progress is real information; it is not a transfer of credit.
A sixty-second check on any NCT number
- Open the record itself, not a summary. Vendor pages and aggregators strip the fields that matter.
- Read the brief summary's first sentence. Self-declaring example records usually say so there or in the title.
- Check the sponsor's other registrations. A program has a history: earlier phases, completions, results. A single orphan record with no lineage is a question, not an answer.
- Check sites and contacts. Many records sharing one site and one email address is a batch, not a portfolio.
- Check the study ID format and compound code. Codes like LY-, and internal identifiers in a large sponsor's house format, belong to that sponsor.
- Check status against dates. "Recruiting" with no update since first posting means nothing has been reported since submission.
- Ask whether results exist. Completed with no posted results and no publication is close to no information at all.
FAQ
Does this mean BPC-157 has no human data? It has a little, and it is old and hard to retrieve — the Pliva PL 14736 work, a Phase 1 and a Phase 2 in ulcerative colitis, neither published as a full peer-reviewed report in the indexed literature — plus a two-participant IV safety pilot in 2025. What it does not have, as of today, is a verifiable ongoing efficacy trial. We said otherwise yesterday.
Why cite registrations at all, if they prove so little? Because they are the only public view of what is about to be tested, and because a program's registry trail is often the most honest summary of it — including the withdrawn, terminated and never-reported entries that no vendor page mentions. The failure mode is not citing them; it is citing them as though they were results.
Should a certificate of analysis be checked the same way? The habit transfers exactly. A number that looks official is a claim by whoever produced it until you check what generated it — which is the same argument we make about reading the raw chromatogram and mass spectrum rather than the summary table on a COA. More on that in quality, and the full compound index is in the library.
This article is educational and for the laboratory research community. Trulogic Labs products are sold for laboratory and research use only and are not for human consumption.