FDA could not find one human study of TB-500. The trials people quote tested a different molecule.

Why this note exists

Two things happened to TB-500 in 2026, and they point in opposite directions. In May, FDA scientists finished an evaluation that could not find a single human study of the compound and recommended against letting compounding pharmacies use it. In July, FDA's own advisory committee voted the other way.[1][3] Between those two dates, the first interventional trial of TB-500 in humans quietly appeared on the registry.[17]

In the index generated on August 15, 2026, 98 of 112 vendors carried a TB-500 listing — 183 listings, eighth of 83 compounds by vendor coverage, behind GHK-Cu, BPC-157 , MOTS-c , NAD+, Selank , Ipamorelin and KPV .[22] That is one of the most widely stocked compounds pepmg tracks, resting on a human evidence base that a federal review team, searching PubMed, Embase, ClinicalTrials.gov, DailyMed, Drugs@FDA, FAERS and the FDA food-and-supplement complaint system, put at zero.[1]

A note on what kind of evidence this is: this compound has almost no human record of its own, so most of what follows is animal, in vitro or about a different molecule — and every paragraph says which. Where a human trial is described, it is a human trial of full-length thymosin β4, and the sentence says so. Registered trials with no reported results are called unreported, not negative.

The molecule

What TB-500 actually is, in FDA's words

FDA's May 15, 2026 memorandum to the Pharmacy Compounding Advisory Committee opens with an identity problem. "TB-500 (free base) is understood today to be a seven amino acid synthetic fragment of thymosin beta-4 (β4) from amino acids 17 to 23 and an acetyl group at the N-terminal leucine amino acid (Ac-Leu-Lys-Lys-Thr-Glu-Thr-Gln-OH). However, TB-500 is a common name and not a United States Adopted Name."[1] The review gives the molecular formula as C₃₈H₆₈N₁₀O₁₄, molecular weight 889.01 g/mol, CAS 885340-08-9, citing PubChem.[1][6]

The naming problem is not cosmetic. FDA writes that it "has encountered multiple salts, and derivatives, including different active moieties, sold commercially under the same common name," and that such conventions "represent a safety risk for patients as they may be dosed with a different BDS than the physician ordered."[1] The free base and the acetate salt are treated as two distinct bulk drug substances throughout.[1]

Identity and search findings as stated in FDA's evaluation of TB-500-related bulk drug substances, dated May 15, 2026.[1]

The biology behind the fragment is a mapping exercise, and it is honest about its own limits. The activities attributed to different stretches of thymosin β4 come from work by Sosne and colleagues that FDA reproduces as figures: the 17–23 segment carries the six-residue actin-binding region LKKTET, and is where actin binding, angiogenesis, hair growth and dermal wound healing are placed on the map.[11][1] FDA's summary of that literature ends on the open question rather than the claim: "[i]t remains to be determined whether actin binding contributes to the ability of the heptapeptide LKKTETQ to stimulate angiogenesis, induce mast cell degranulation, increase angiogenesis, and promote wound healing."[1]

One published analysis has actually looked inside a commercial product. A Ghent University doping-control laboratory reported in 2012 that it detected and identified Ac-LKKTETQ in TB-500 by high-resolution mass spectrometry, then synthesised the peptide independently to build a detection assay for plasma and urine.[7] That is a single laboratory characterising the formulation it obtained, fourteen years ago. It establishes what was in that material; it says nothing about what is in any vial sold today, and pepmg makes no claim about the contents of any vendor's product.

The regulatory event

FDA's reviewers said no. The committee said yes.

The nomination itself was a mess, and FDA says so. It came from Wells Pharmacy Network, proposing a 3 mg/mL lyophilized powder for subcutaneous and intramuscular injection, for wound healing.[1] FDA lists four inconsistencies in the package: the nominated substance was the free base while the accompanying certificate of analysis was for the acetate; the CoA's CAS number, molecular formula and molecular weight were for the acetate; the molecular formula supplied "does not match with either TB-500 (free base) or TB-500 acetate"; and the "AKA CAS #" given "does not match with either."[1] The nomination was then withdrawn, and FDA carried on evaluating both substances at its own discretion.[1]

The conclusion is unambiguous. "On balance, the physiochemical characterization, information on historical use, lack of any evidence of effectiveness, and safety information for both TB-500 (free base) and TB-500 acetate weigh against their being added to the 503A Bulks List … Accordingly, we propose not adding TB-500 (free base) or TB-500 acetate to the 503A Bulks List."[1]

At the July 23–24 meeting the committee voted to recommend TB-500-related substances for the list. STAT reported the tally as 8 to 6, and reported that "FDA staff told the panelists they recommended against giving compounders the green light due to a lack of clinical evidence demonstrating safety and efficacy."[3][4] FDA had not posted final summary minutes for the meeting as of August 16, 2026, so the number above is a press report rather than a primary record.[2][3]

What the vote is not: an approval, a finding of efficacy, or an immediate change in the law. Advisory committee recommendations are non-binding, inclusion on the 503A Bulks List proceeds by rulemaking, and pepmg set out the mechanics in its note on the July votes.[5][21] The compound remains unapproved: FDA records that there is no USP or National Formulary monograph for either substance, that neither "is a component of an FDA-approved drug," that TB-500 is not recognised in the European or Japanese Pharmacopoeias, and that no approved TB-500 product exists in Canada, Australia, the UK, Belgium, Ireland, France, Norway, Germany, Spain or Italy, nor any EMA authorisation.[1]

The gap

What a federal review team found when it went looking for humans

The searches are worth reciting because their breadth is the finding. For historical use FDA searched PubMed, Embase, GlobalEdge, NatMed Pro, USP-NF, the FAERS public dashboard, Google, Compounding Today, the European and Japanese Pharmacopoeias and the Outsourcing Facility Product Reporting Database; the result was that "[a]fter conducting a literature search, no articles were found in which TB-500 was administered to humans."[1] No outsourcing facility had reported compounding it, and no pharmacy was found marketing a compounded TB-500 product.[1]

For effectiveness: "[w]e performed our own search of published medical literature … however, we did not identify data, such as clinical studies, on the use of TB-500 in humans."[1] The three references the nominator submitted "are studies on Thymosin β4," and none discussed a compounded TB-500 formulation.[1]

For safety: FDA's Office of Surveillance and Epidemiology searched FAERS through March 26, 2025 and "did not retrieve any reports."[1] A search of the human-foods complaint system from January 2004 to March 2025 returned two cases about "blended TB-500 and BPC-157," neither including a safety assessment.[1] No case reports in the published literature. The section closes: "potential safety risks associated with the use of TB-500 in humans are unknown."[1]

An absence of adverse-event reports is not a safety finding. FDA says as much in a footnote about FAERS — reporting is voluntary, the agency does not receive all reports, and it "cannot make definitive conclusions regarding the safety of TB-500."[1] A compound with no reported harms and no reported exposures has been studied, not exonerated.

Animal and in vitro data, labeled as such

Even the preclinical file mostly tested something else

The wound-healing citation that circulates under TB-500's name is Philp and colleagues, 2003. Four 3-mm punch wounds were made in 26-month-old female BALB/cBy mice; 50 µL of thymosin β4, of the non-acetylated heptapeptide LKKTETQ, or of vehicle was applied topically on the day of wounding and 48 hours later. Histology at day 7 suggested that either thymosin β4 or the non-acetylated heptapeptide significantly increased epidermal closure and collagen content.[9][1] Aged mice, topical application, a peptide that is not TB-500. FDA adds the reason the difference matters: acetylation "irreversibly alters their charge, hydrophobicity, and size," so the non-acetylated profile "cannot be directly extrapolated" to acetylated TB-500.[1]

When someone did test the acetylated compound, the result went the other way. In a 2024 in-vitro study, confluent fibroblast cultures with scratch-generated wounds were incubated for eight hours with TB-500 free base at 50 µg/mL or vehicle; wound closure "was not significantly different" between them.[10][1] A metabolite, N-acetylated LKKTE, produced a small but significant closure at the same concentration.[10] FDA notes the study lacks a concentration-response analysis, and that it remains unknown whether TB-500 has wound-healing properties in vivo at all.[1]

The pharmacokinetics come from horses. In thoroughbred geldings given 10 mg subcutaneously, plasma TB-500 peaked at 0.05 and 0.08 ng/mL between 60 and 120 minutes and became unquantifiable between 6 and 10 hours, and the peptide was recovered as a ladder of C-terminally truncated metabolites.[8][1] A veterinary preparation of TB-500 appeared in 2011 and was marketed to boost performance in equine and greyhound sport.[8][1] Human tissue enters this literature only as microsomes and serum in a test tube, in metabolite-profiling work.[10][23]

What is missing is as informative as what is there. FDA states that it identified no acute toxicity, no repeat-dose toxicity, no genotoxicity, no developmental or reproductive toxicity and no carcinogenicity studies of TB-500 free base or TB-500 acetate, and no in-vivo pharmacology study of TB-500 itself for wound healing.[1] It also flags an untested risk specific to the proposed use: peptides given by injectable routes "may pose a significant risk for immunogenicity, potentially amplified by aggregation," and no study of immunogenicity or aggregation for TB-500 was submitted or found.[1]

The substitution

The human trials that get borrowed, and what they actually tested

Thymosin β4 — the 43-amino-acid protein — does have a human record. It is small, mixed, and almost entirely by routes nobody in the research market uses. Reading it as evidence for a subcutaneously injected heptapeptide requires two substitutions at once: a different molecule and a different route.

Human trials of full-length recombinant or synthetic thymosin β4, not of TB-500. Enrollment and results as reported in each publication.

Take them one at a time. The 2010 phase 1 gave synthetic thymosin β4 intravenously to four cohorts of ten healthy subjects at 42, 140, 420 or 1260 mg, first as a single dose and then daily for 14 days; adverse events were infrequent and mild to moderate, with no dose-limiting toxicities or serious adverse events.[12] That is a genuine human safety signal — for an intravenous protein at gram-scale doses, which is not what a 10 mg vial of a heptapeptide is.

The eye programme is where the strongest claims come from, and it is where the numbers deserve the most care. A 2015 phase 2 in severe dry eye randomised nine patients and reported significant improvements in discomfort and corneal staining at day 56.[13] The 2022 phase 3 in neurotrophic keratopathy enrolled 18: complete healing at four weeks in 6 of 10 treated versus 1 of 8 on placebo, which the authors describe as "a strong efficacy trend" at p = 0.0656 — that is, the primary comparison did not reach significance, and several secondary findings did.[14] Both were sponsored by the developer, with several authors employed by it, as the papers disclose.[13][14]

An independent body has since graded that trial. The Cochrane review of interventions for neurotrophic keratopathy, published December 2025, included seven randomised trials totalling 494 participants and assessed the 0.1% RGN-259 trial as finding that it "may not increase the proportion of participants with corneal re-epithelialization" — risk ratio 9.00 with a 95% confidence interval running from 0.57 to 141.88 across 18 participants, rated low-certainty evidence.[15] A confidence interval that wide is what eighteen people buys.

The cardiac trial is the newest and the most carefully reported. Published in Cardiovascular Research in December 2025, it is mostly a mouse and cell study; its clinical component was "a randomized, placebo-controlled, double-blind trial involving 96 STEMI patients." The authors report that infarcted areas were significantly reduced in the subgroup that received the first dose within 8 hours after percutaneous coronary intervention (n = 43), and state plainly that "the overall differences in infarcted areas were not significant[] between the rhTB4 group and the placebo group (n = 96)."[16] Their own conclusion calls for "[f]urther rigorous randomized studies."[16] Three authors are employees of the manufacturer.[16] A significant subgroup inside a trial whose overall comparison missed is a hypothesis, not a result.

A registry search on August 16, 2026 returned 18 studies with thymosin beta 4 as an intervention, against one for TB-500.[18] The literature is real. It belongs to the other molecule.

The new fact

There is now exactly one registered human trial of TB-500

It was first posted on March 23, 2026 — after FDA's database searches, before the memo that reported none — and FDA's evaluation does not mention it.[17][1] That is a timing observation, not a criticism: the agency's cited access dates for these databases run from February 2024 to April 2025.[1]

NCT07487363, acronym TBRIDGE-CV, is sponsored by Hudson Biotech: "A Phase 1/2, Randomized, Double-Blind, Placebo-Controlled, Sequential Dose-Escalation Study of TB-500 (Thymosin Beta 4 17-23 Fragment) in Adults With Stable Atherosclerotic Cardiovascular Disease to Evaluate Safety, Tolerability, Pharmacokinetics, and Exploratory Cardiovascular Biomarkers."[17] Estimated enrollment 80, three sequential dose cohorts randomised 3:1 against placebo, actual start February 5, 2026, estimated primary completion February 14, 2027.[17]

Two details are worth holding onto. First, the primary outcomes are the incidence of treatment-emergent and serious adverse events — this is a safety and tolerability study. Flow-mediated dilation, hs-CRP and NT-proBNP are secondary, and pulse wave velocity is explicitly exploratory and only "if available at site."[17] Nothing in it measures wound healing, the use the compound is sold and was nominated for.[17][1] Second, the registry's masking field says single (participant), while the free-text masking description says investigators and outcome assessors also remain blinded; the structured field is the one that indexes, and the two do not agree.[17]

The record shows no results, and the trial is recruiting.[17] Anyone quoting it before February 2027 is quoting a protocol.

The market

Sixty-four listings put the protein's name on the fragment's vial

This is the part pepmg can measure directly, because it is what vendors publish. In the index generated August 15, 2026, TB-500 appeared in 183 listings across 98 of 112 vendors — 161 single-compound listings and 22 blends, most commonly BPC-157 pairings.[22] Where a size is stated, 10 mg is the standard vial: 85 of the 151 solo listings with a size.[22]

The naming splits almost exactly against FDA's warning. 64 of the 183 listings, from 30 vendors, put "thymosin beta-4," "TB4" or "Tβ4" in the product title.[22] Several go further and print an explicit length — product names in the index include "TB-500 (Thymosin Beta 4) (43 AA)" and "TB-4 Full Chain," which assert the 43-amino-acid protein on a product sold under the fragment's name.[22] Against that, 8 listings from 7 vendors flag the fragment at all, with titles such as "TB-500 Frag 17-23" or "TB-500 Fragment (17-23)."[22] Exactly one listing does both.[22]

These counts describe product titles, not molecules. pepmg has no analytical access to any vendor's vial and makes no claim about what any of them contains; a title saying "43 AA" is a claim its vendor published, and a title saying nothing is not evidence of anything either way. What the counts do establish is that the conflation FDA identified in the literature is also the default in the market — the same interchangeable use, on the labels people buy from.[1][22]

One consequence is worth stating for anyone in tested sport. The 2026 World Anti-Doping Agency Prohibited List bans, under S2.3, "Thymosin- β4 and its derivatives e.g. TB-500" — prohibited at all times, in and out of competition.[19][20] WADA groups the two where FDA distinguishes them; both are correct for their own purpose, and neither reading gets an athlete off.

Three things this note is not saying

First, it is not saying TB-500 does not work. It is saying that the question has not been asked in humans. FDA searched and found no clinical studies, no pharmacokinetic or pharmacodynamic data, no case reports and no adverse-event reports; absence of evidence in a compound this widely sold is a finding about the literature, not a verdict on the molecule.[1]

Second, it is not saying thymosin β4 is a failed drug. Its human record is small and mixed rather than negative: a clean 40-person intravenous phase 1, a 9-person phase 2 and an 18-person phase 3 in ophthalmology that Cochrane rates low-certainty, and a 96-patient cardiac trial whose overall endpoint missed while a prespecified early-dosing subgroup did not.[12][13][14][15][16] Those are real trials in real patients, and they are ongoing science.

Third, it is not saying the advisory committee was wrong. Compounding access is a policy judgment about patients, prescribers and unmet need, and reasonable people weigh it differently from a reviewer scoring a data package. What can be said without taking a side is that FDA's scientists and FDA's advisers read the same file and reached opposite recommendations, and that the file itself contains no human data.[1][3]

What it is saying is narrower, and it is a labelling point before it is a science point: the seven-amino-acid compound in the vial and the 43-amino-acid protein in the citations are different substances, FDA has written that sentence down, and most of the listings in this market still use the two names as though they were one.[1][22]

Questions people are asking