BPC-157 and TB-500 — what the recovery peptide literature actually contains

BPC-157 and TB-500 sit next to each other on the recovery shelf of this catalogue, they get discussed in the same sentences, and they have been given together in exactly one published experiment. Neither holds a marketing authorisation in any jurisdiction. Both are named on the World Anti-Doping Agency Prohibited List in force from 1 January 2026 (24).

This post carries no figure row and no chart. That is not an oversight and it is not modesty. Neither compound has a controlled result in people that a figure could honestly be lifted from, and the most useful thing a reader can be told about the two is the shape of the hole where that result would sit.

The holes are different shapes. BPC-157 has a large literature and a narrow one: the great majority of it comes from a single research group in Zagreb, and independent replication is sparse (9). TB-500 has a small literature, and the impressive human results usually quoted under its name were produced by a different, longer molecule — full-length Thymosin Beta-4 — which the fragment is not (12, 13).

Everything below describes what published studies measured, in the models they used. No trial has compared these two compounds in people, and none of this is guidance for use in people.

BPC-157 is a synthetic peptide of fifteen amino acids, GEPPPGKPADDAGLV, described in the published work as a partial sequence of a larger protein isolated from human gastric juice (1, 2). Publications carrying the name have accumulated for three decades. A systematic review of the musculoskeletal literature screened 544 records published between 1993 and 2024 and included 36 studies: 35 preclinical and one clinical. It graded the whole body of work as level IV and level V evidence, and recorded that no clinical safety data were found (6).

That is the first limitation, and it is the one most readers already know about. The second is less visible and matters at least as much. In a 2025 exchange in the literature, the authors of a review of the compound reported that a PubMed search on 20 May 2025 returned more than 190 articles containing the term BPC 157, and that more than 80% of them listed P. Sikiric or S. Seiwerth as first or senior author — the group at the University of Zagreb on whose work the field largely rests. Independent groups, they wrote, have contributed only a handful of in vitro or short-term rodent studies (9).

Their conclusion deserves stating in their own terms: heavy reliance on self-replication restricts how far findings generalise and increases the risk of confirmation bias (9). None of that makes any individual reported finding wrong. It means the findings have largely not been checked by anyone else, and that a count of publications is not a count of independent tests.

The concentration has a concrete consequence, which the same authors name. Experiments in this literature routinely use a single dose level, so there is no information about what happens at higher, repeated or long-term exposures, and no way to tell whether the reported effects are dose-dependent or saturable (9). A second group repeating an experiment tends to vary exactly those things, because it has no stake in the original result. That is what independent replication buys, and it is what is missing here.

One unresolved problem runs through the reviews. Two of them report a plasma half-life under 30 minutes, with metabolism in the liver and clearance by the kidneys, against effects described in animals as lasting hours to days (6, 7). A 2026 review of the compound's pharmaceutical development states the position without hedging: no approved formulation, no validated dosing regimen, no completed Phase 2 trial, and available clinical data drawn from fewer than thirty subjects across three uncontrolled pilot studies, none of which used a standardised preparation (7).

Thymosin Beta-4 is a protein of 43 amino acids, present in most mammalian cells, where it binds monomeric actin one to one and holds it out of the polymerising pool. Safer and colleagues settled its identity in 1991, showing that the actin-sequestering peptide isolated from human platelets and the thymic peptide already known as thymosin beta 4 were the same molecule (11, 13).

TB-500 is not that molecule. When Esposito and colleagues took a TB-500 preparation apart by high-resolution mass spectrometry, what they identified was a seven-residue piece of it — Ac-LKKTETQ, positions 17 to 23 of the parent, with the amino terminus artificially acetylated (12). The equine doping-control literature describes the same segment the same way (14).

The stretch is not an arbitrary one. Sosne and colleagues mapped the parent protein's activities onto short sequences and placed angiogenesis, wound healing and cell migration on LKKTETQ plus one adjacent residue — the central actin-binding domain — while a separate amino-terminal fragment, Ac-SDKP, carried the anti-inflammatory and anti-fibrotic activity (13). Carrying the right binding site is a reason to expect something. It is not a demonstration that anything follows.

The distance between the two shows up most clearly in who has been given what. Full-length Thymosin Beta-4 has been through trials in people. Four cohorts of ten healthy volunteers each received a single intravenous dose of 42, 140, 420 or 1,260 mg, then the same dose daily for fourteen days; the report describes infrequent adverse events, mild or moderate in intensity, with no dose-limiting toxicity (17). Under the name RGN-259 the protein was taken into eye disease: a Phase 2 trial in nine patients with severe dry eye reported improvements in discomfort and corneal staining against vehicle (18), and a Phase 3 trial in neurotrophic keratopathy reported complete healing of the epithelial defect at four weeks in 6 of 10 treated patients against 1 of 8 on placebo, which the authors describe as a strong efficacy trend rather than a significant difference, at p = 0.0656 (19). Two further Phase 3 dry eye trials, ARISE-2 and ARISE-3, enrolled 601 and 700 participants and are recorded as completed (20).

None of it was done with TB-500. The fragment has one registered human trial: an industry-sponsored Phase 1/2 dose-escalation study in adults with stable atherosclerotic cardiovascular disease, with eighty participants planned, recruiting since February 2026 and due to reach primary completion in February 2027. It has posted no results (16).

The direct literature on the fragment is otherwise analytical. Ho and colleagues built a doping-control assay and confirmed the peptide and its metabolites in urine and plasma from horses after a single administration (14). Rahaman and colleagues followed it through human serum, enzyme preparations and rat urine; they reported Ac-LK as the metabolite present at the highest concentration early on, with Ac-LKK still detectable at 72 hours, and found that the wound-healing activity in their cell assays sat with a longer metabolite, Ac-LKKTE. They state the position for the fragment itself without hedging — the biological effects of TB-500 have not been documented — and raise the possibility that activity previously attributed to it belongs to a metabolite rather than to the intact peptide (15).

Only one published study gave both compounds under the same protocol, and it is worth describing exactly. Thirty-two male Sprague-Dawley rats had the Achilles tendon transected and repaired, then were assigned to four groups of eight — control, BPC-157, TB-500, or the two combined — and treated intraperitoneally for four weeks (23).

Both peptides were associated with improved histopathological parameters and extracellular matrix organisation. On maximum load to failure, both treated groups sat above controls, but only the TB-500 group reached statistical significance; the BPC-157 group's total scores were numerically lower than controls without reaching significance. Combining the two conferred no additional benefit over either alone (23).

The authors call the study exploratory and its findings preliminary, and note that dose optimisation and longer-term work remain to be done (23). Eight animals per group is a small experiment, it is in rats, and it is one study. It is also the only place in the literature where these two compounds have been set against each other under the same conditions — which is why the table below is otherwise made of descriptions rather than results.

Set side by side, the two compounds come up short in opposite directions, and neither shortfall is the one people usually look for. BPC-157 has volume: three decades of publications, a systematic review, proposed mechanisms in four separate signalling systems. What it does not have is a second set of hands. TB-500 has the reverse problem. There is a real programme in people behind the name, with Phase 1, Phase 2 and Phase 3 trials in it — and it belongs to the whole protein rather than to the seven-residue piece of it that a TB-500 preparation was found to contain.

So the question worth asking is different for each. Of BPC-157: who else has seen this? Of TB-500: which molecule was administered? Both are answerable from the abstracts, and both are usually skipped.

There is a third habit worth naming, because it flatters both compounds at once. Neither has a public failure behind it — no missed primary endpoint, no discontinued programme, nothing of that kind. That reads as a clean record, and it is not one. An untested compound has no failures because it has no results, and the absence of a bad outcome is not the presence of a good one. It is what a literature looks like before the question has been asked.

Which is why this post ends where it began, with no figures in it. The honest summary of both evidence bases fits into a single sentence: what either compound does in people has not been measured, and everything published so far describes animals, cultured cells, or a different molecule.