Evidence review
Does the published thymosin beta-4 research apply to the compound sold as TB-500?
The answer
No. TB-500 is a seven-residue fragment of a 43-residue protein, and the clinical record belongs to the protein. Searched on its own names, TB-500 returns one record under the clinical-trial publication type, and that record is a study of tuberculosis.
Thymosin beta-4 is a protein of 43 amino acids involved in actin regulation and tissue repair, and it has a real research literature under its own name. TB-500 is a product name attached to material sold on the research-chemical market, and the two are routinely written about as though they were the same substance.
They are not, and the difference has been characterized in print. A 2012 analysis synthesized and identified the substance present in a TB-500 formulation as the N-terminal acetylated 17-23 fragment of thymosin beta-4 — the sequence Ac-LKKTETQ — and proposed a method for detecting it in plasma and urine [1]. Seven residues of forty-three is not a formulation difference. It is a different molecule that shares an origin.
What is in any given vial is a separate question again. A 2023 laboratory analysis of products sold under the TB500 and TB1000 names reported that their contents were not consistently what the product descriptions claimed [2]. That paper is about misbranded and adulterated material sold over the internet, which is the supply channel this compound has, because it has no other. The research-use-only entry sets out what that labeling posture does and does not mean.
The two names were searched separately in August 2026 and their counts were never added together. Thymosin beta-4, across four spellings, returns 1,107 indexed records, of which 598 carry the Humans subject heading; under the clinical-trial publication type it returns 8, and under the stricter randomized-controlled-trial type it returns 7 [3].
TB-500, across its two spellings, returns 31 indexed records, of which 21 carry the Humans subject heading. Under the clinical-trial publication type it returns 1. Under the randomized-controlled-trial type it returns 0 [3].
The filter was proved live in the same run against a control term that returned 8,904 records under the clinical-trial type and 6,113 under the randomized type; the invalid publication-type tag this index was once burned by returned 0 for that same control term, and a deliberately invalid record identifier errored in the same session [3]. A dead filter and a true absence print the same number, so the control is the part that makes the 0 worth reading.
It is not one. The single record returned under the clinical-trial publication type is a logistic regression analysis of environmental variables and the incidence of tuberculosis in respiratory patients, which matched the search on the letters TB [3].
It was found by reading the hit rather than counting it, and it is worth naming as a method point rather than a curiosity. A positive control proves that a filter is alive. It says nothing about whether the records the filter returned are about the subject. The honest count for this compound is zero human trials, and reaching it required opening the one result that stood between the search and that answer.
The 21 records carrying the Humans subject heading do not close the gap either, because most of them are analytical chemistry written for sport drug testing rather than for therapy: extraction methods, mass-spectrometry screens, and metabolism studies whose purpose is to detect the substance in an athlete's urine [3]. That is a literature about finding TB-500 in people, not about what it does to them.
The protein's clinical record is small but it is real, and it belongs to specific indications. A first-in-human phase 1 study gave recombinant human thymosin beta-4 to healthy volunteers by intravenous injection in seven single-dose cohorts of 54 subjects in total and three multiple-dose cohorts of 30, reporting mild to moderate adverse events, no dose-limiting toxicities, and dose-proportional exposure [4]. That is a safety and pharmacokinetic result in healthy people, and the authors framed it as groundwork for a trial in acute myocardial infarction.
A phase 2 randomized trial reported improvement in the signs and symptoms of severe dry eye [5]. Those are the indications the protein has been carried into: the eye and the heart, by intravenous or ophthalmic administration, in trials run by sponsors who manufactured a defined product.
None of that is a record about tendon repair, ligament injury, or recovery from training, which are the uses the fragment is sold for. Nor does a trial of a manufactured 43-residue protein establish anything about a seven-residue fragment obtained from an unlicensed supplier. The repair peptides class holds two neighbors with the same shape of problem: BPC-157, whose human trial record is empty, and GHK-Cu, whose evidence is topical and whose market is injectable.
TB-500 was removed from Category 2 of the interim 503A bulks list in April 2026 without being added to Category 1, and in July 2026 the Pharmacy Compounding Advisory Committee voted to recommend adding it to the section 503A list. The margin is reported as 8–6 with one abstention by McDermott Will & Schulte and by LumaLex Law, whose separate accounts of the meeting agree substance by substance [6, 7]; the FDA has published no minutes, transcript or vote summary for it, so that figure is secondary and this review says so where it prints one. The vote is advisory, no FDA final rule has issued, and no licensed United States pharmacy may lawfully compound it. The monograph carries that ledger with the date each entry was verified, the vote ledger carries the arithmetic of the whole meeting and the sourcing behind it, and the pending page explains what an advisory recommendation changes, which is nothing yet.
That posture and the evidence above answer two different questions, and conflating them is the second mistake this page exists to prevent. A committee recommending a substance for a compounding list is not a committee finding that the substance works. The vote concerned whether pharmacies should be permitted to prepare it; the trial record concerns whether anyone has tested it. On the first question there is a recommendation and no rule. On the second, for the fragment as sold, there is nothing at all.
Monographs
Reference tables
Every source below is inherited from a monograph in this index, where it was verified at writing time; this review introduces none of its own. The pointer beside each entry names the monograph and the reference number it came from.