Thymosin Alpha-1

What Thymosin Alpha-1 is and how it works

Thymosin Alpha-1 is a peptide of 28 amino acids. It is not a designed molecule: it is the first 28 residues of prothymosin alpha, a much larger protein involved in chromatin remodelling, released by a lysosomal asparaginyl endopeptidase identified as mammalian legumain, which cleaves the Asn28–Gly29 bond. That paper matters because it settled an old argument. The peptide had been assumed to be an artefact of the way thymic extracts were prepared, and the finding that it is the main product detected in vivo, free in the cytosol at concentrations similar to the parent protein, established that the body makes it.^[3] The material used in clinical trials is a synthetic copy, and its international non-proprietary name is thymalfasin; a randomized Japanese trial report gives the same 28-amino-acid description of the substance it administered.^[6]

This is not TB-500, and it is not Thymosin Beta-4. The confusion is historical rather than careless. Both molecules were pulled out of the same crude preparation — "thymosin fraction 5", a mixture of polypeptides — and thymosin alpha 1, polypeptide beta 1 and thymosin beta 4 were each isolated from it and tested separately. A review of that history is blunt about what followed: none of the isolated peptides turned out to be a thymic hormone, polypeptide beta 1 is truncated ubiquitin, and thymosin beta 4 is the main intracellular G-actin-sequestering peptide of most vertebrate cells rather than an immune signal.^[8] Thymosin beta 4 and its fragment TB-500 have their own page on this site, their own separate literature and, as the regulatory section below records, their own separate entry in FDA's own compounding documents. None of that evidence is evidence about this peptide, and there is no "thymosin class effect" that would let a result travel between them in either direction. Thymalin — a different thymus-derived preparation with a literature of its own — is not thymalfasin either, and neither are thymopentin, thymulin or thymic humoral factor.

The proposed mechanism is immunomodulation through dendritic cells. In mice and in culture the peptide induced maturation and interleukin-12 production by fungus-pulsed dendritic cells through a p38/NF-κB route, signalling by way of MyD88 and distinct Toll-like receptors; in vivo it activated Th1-dependent antifungal immunity and protected transplanted mice from aspergillosis.^[5] That is a mouse and cell-culture result, and the word usually attached to it — "booster" — does not survive contact with the human-cell work. In human monocyte-derived dendritic cells the same peptide raised HLA class I and II expression and IL-6, TNF-α and IL-8 secretion under viral TLR3 and TLR7/8 agonists, and under bacterial TLR2 and TLR4 stimulation and BCG infection it drastically lowered the same measurements.^[11] Whatever it does, it does not do one thing.

What the research actually shows

Sepsis, where the largest trial missed. The TESTS trial randomized 1,106 adults with sepsis across 22 Chinese centres to subcutaneous peptide or placebo every 12 hours for seven days. In the 1,089-patient modified intention-to-treat analysis, 28-day all-cause mortality was 23.4% against 24.1% — hazard ratio 0.99, 95% CI 0.77 to 1.27, P=0.93 — and no secondary or safety outcome differed significantly. The authors' own conclusion is that the trial found no clear evidence that the peptide decreases 28-day all-cause mortality in adults with sepsis. Prespecified subgroup analyses showed interactions by age and by diabetes, and a subgroup interaction in a trial whose primary endpoint was not met is a hypothesis, not a finding.^[17] The earlier ETASS trial randomized 361 patients in six Chinese hospitals; 28-day mortality was 26.0% against 35.0%, but the nonstratified analysis gave P=0.062 and the relative risk of 0.74 carried a confidence interval of 0.54 to 1.02 that crossed 1. Its log-rank P of 0.049 is reported alongside, and it is not the same thing as the nonstratified result. ETASS was also single-blind and used a control group receiving conventional therapy rather than a placebo.^[10] A 2025 meta-analysis of 11 randomized trials found a pooled odds ratio of 0.73 (0.59–0.90) — and then found no mortality benefit in the high-quality subgroup (0.82, 0.65–1.03) or the multicentre subgroup (0.86, 0.68–1.08), with a trial sequential analysis concluding the accrued sample is inadequate.^[18]

Hepatitis B, where the results are real but small. In 98 Taiwanese patients, complete virological response at 18 months was 40.6% after a 26-week course against 9.4% in untreated controls (P=.004); the 52-week arm reached 26.5% and did not separate from control (P=.068), and response rates were similar across all three groups when first assessed at the end of therapy.^[2] A 316-patient Japanese trial compared 0.8 mg against 1.6 mg for 24 weeks and found similar efficacy rates in both arms overall — while reporting that patients with advanced fibrosis achieved a significantly better response rate on 1.6 mg than on 0.8 mg, a result the authors themselves qualify as an intragroup analysis in a trial that did not stratify patients by liver biopsy.^[6] A 62-patient trial run against interferon alpha cuts in both directions, and both halves belong here. Its composite complete response — ALT normalization together with loss of HBV DNA and of HBeAg — did not differ significantly between the two drugs at either timepoint: 31.0% against 45.5% at the end of treatment, and 48.3% against 27.3% after six months of follow-up, P>0.05 both times, with significance reached only against a separate historical control group. The individual endpoints ran the other way. By the end of follow-up the proportions achieving ALT normalization and negative HBV DNA were significantly higher on the peptide than on interferon alpha and than in the historical controls, at three-fold higher odds against interferon. The trial nonetheless closed with the authors writing that a response rate of 48.3% is still less ideal.^[7]

Cirrhosis, where the hard endpoint missed. 690 patients with HBV-related compensated cirrhosis were randomized to entecavir with the peptide or entecavir alone for 52 weeks and followed a median 38.2 months. The primary endpoint was a composite of liver decompensation, hepatocellular carcinoma or death, and its cumulative incidence was similar between groups; virologic, serologic and biochemical responses at week 104 were similar too. The authors' closing sentence converts that into a tendency to inhibit cancer development, which is a description of a non-significant direction rather than a second result.^[12] A meta-analysis of seven randomized trials in the same disease found higher response rates at 24 weeks that were no longer significantly different at 48 and 52 weeks, and recorded that every included patient came from mainland China.^[13]

Hepatitis C, where monotherapy failed outright. Nineteen Italian patients were randomized double-blind to six months of peptide or placebo. There were no significant changes in mean serum ALT in either group, no patient cleared HCV-RNA, and the authors concluded that the regimen is not effective.^[1] Added to interferon in 41 patients, it produced a significantly better end-of-treatment virological response (P=0.03), while the sustained biochemical and virological response six months later was not different between the groups.^[4]

COVID-19, where the largest dataset points the other way. A multicentre retrospective cohort compared 306 treated patients with 1,976 untreated ones in Hubei. Every crude outcome was worse in the treated group, and after adjustment use remained associated with a higher non-recovery rate (odds ratio 1.5, 95% CI 1.1–2.1, p=0.028), rising to 5.4 (2.1–14.0) among patients admitted to intensive care.^[14] A meta-analysis of eight studies reported lower mortality (risk ratio 0.59, 95% CI 0.37–0.93) with I²=84%, no difference in the need for mechanical ventilation or in hospital length of stay, and a meta-regression finding the mortality benefit significantly affected only by sample size and by sex.^[16]

Vaccine response and cancer surgery, both thinner than they look. In 94 hemodialysis patients given an adjuvanted pandemic H1N1 vaccine, the report states that the CHMP licensing criteria were fully met in the two peptide groups. It does not state whether the vaccine-only group met them, and this page does not supply that half by inference. What the report does say about the vaccine-only group is that by Day 42 the decline in geometric mean haemagglutination-inhibition titre was greater in the peptide groups than in it, so whatever boost there was also faded faster.^[9] A propensity-matched retrospective analysis of 468 patients after resection of solitary HBV-related hepatocellular carcinoma reported a multivariate hazard ratio of 0.308 for overall survival. That is a chart-derived comparison, and the randomized trial whose composite endpoint included hepatocellular carcinoma found no difference between groups.^[15][12]

Where the evidence is weak

The claim most readers arrive with has no trial at all. A PubMed search for (thymalfasin OR "thymosin alpha 1") AND (immunosenescence OR "immune support" OR "immune enhancement") AND healthy on 1 September 2026 returned no records, and a PubMed search the same day for (thymosin alpha 1 OR thymalfasin) AND (healthy volunteers OR healthy adults) AND randomized returned three records, none of which randomized healthy participants to the peptide — one severe-sepsis trial, one cytokine study in chronic hepatitis B patients, and one lung-cancer immunology paper in which healthy subjects appear only as an age-matched comparison group. A PubMed search on the same date for (thymalfasin OR "thymosin alpha 1") AND (long-term OR chronic administration) AND (healthy volunteers) returned no records either. Every number on this page was generated in people who were acutely or chronically ill.

The headline evidence is negative, and reading it any other way requires ignoring the primary endpoints. The newest and by far the largest randomized trial missed.^[17] The largest hepatology trial missed on a hard clinical composite.^[12] Monotherapy in hepatitis C failed outright.^[1] The sepsis meta-analysis that reports a pooled benefit loses it in exactly the subgroups built to be more trustworthy.^[18] A page that assembled these same sources and came out enthusiastic would have misread them.

The trial base is geographically and institutionally concentrated. The phase 3 sepsis trial ran across 22 centres in China;^[17] its predecessor across six Chinese teaching hospitals;^[10] the retrospective COVID cohort across five hospitals in Hubei province;^[14] and the meta-analysis of cirrhosis trials records in its own abstract that every included patient came from mainland China.^[13] The dose-comparison trial studied Japanese patients^[6] and the two small hepatitis C trials Italian ones.^[1][4] None of the eighteen sources cited on this page reports a trial conducted in North America. The 2025 sepsis meta-analysis was written by authors who also appear on the author lists of the two trials that supply most of its patients,^[18][17][10] so it is not an independent appraisal of that evidence, and the phase 3 trial itself declares support from its commercial sponsor with twelve of its authors holding grants from that company.^[17] The trial report also carries a published erratum.^[17]

No amount is established, and the one dose-response signal on record is a qualified subgroup. The only head-to-head comparison of two amounts found similar efficacy at 0.8 mg and 1.6 mg across the trial population as a whole, and reported a significantly better response rate at 1.6 mg among the patients with advanced fibrosis — a result its own authors mark as an intragroup analysis in a trial whose patients were not stratified by liver biopsy, which makes it a hypothesis about one subgroup of hepatitis B patients rather than a dose-response curve anyone could read an amount off.^[6] The phase 3 sepsis trial reports its route, its interval and its duration but no milligram figure in its abstract,^[17] so the largest exposure on record cannot be expressed as an amount here at all. Neither figure was administered to a healthy person.

Safety data exist only as a by-product of efficacy trials. A PubMed search for (thymalfasin[ti] OR "thymosin alpha 1"[ti] OR "thymosin alpha-1"[ti]) AND (safety[ti] OR tolerability[ti] OR "adverse events"[ti]) on 1 September 2026 returned seven records — two retrospective oncology series, two study protocols, one review, one randomized hepatitis B trial and one pilot study in HIV — and none of them is a standalone safety or adverse-event study. What is known about tolerability comes from seven-day exposures in intensive care and from 24- to 52-week courses in liver disease, monitored throughout.^[17][6] The one large dataset that looked for harm rather than benefit found an association with worse outcomes.^[14]

Why this page's compound-level grade is the lowest of the three. The badge on this page reads Anecdotal reports only, and that is not a statement that the trial base is anecdotal — it plainly is not, and the sections above describe multi-centre randomized trials of up to 1,106 patients. The compound-level grade on this site is the grade of the typical claim rather than the best one, and the typical claim here is general immune support in a healthy adult, for which the searches above returned nothing to grade. The individual claim grades in the evidence table carry the real picture: randomized human evidence in sepsis, hepatitis B, cirrhosis and hepatitis C, and the lowest grade for everything else. Where the two readings pull apart, this site grades down.^[11]

Legal and regulatory status

Thymosin Alpha-1 is not approved by the FDA for any indication, and there is no approved US product containing it. That was verified at authoring time rather than recalled: Drugs@FDA returned no matches for thymalfasin, thymosin alpha-1 or Zadaxin across brand-name, generic-name and active-ingredient queries, and the approved-labeling database returned none either, while same-session control queries for somatropin, semaglutide and Ozempic returned records as expected.

A reader searching FDA's own databases will nonetheless find nine entries in the National Drug Code directory, and it is worth saying plainly what they are. Every one is a BULK INGREDIENT or BULK INGREDIENT FOR HUMAN PRESCRIPTION COMPOUNDING registration for a powder, with no route of administration and no application number — a manufacturer registering an active pharmaceutical ingredient with the agency. It is not an approval, and it does not become one by being on an FDA server.

Three FDA compounding documents bear on it, and they do not all say the same thing. The peptide appears in none of the three 503A bulk-drug-substance nomination categories: the 503A category list current to 14 May 2026 returns no hits for thymosin, thymalfasin or Zadaxin, while in-document controls for other substances in that file return theirs. It does appear on the equivalent 503B list — the outsourcing-facility document, updated 21 March 2025 — as Thymosin alpha-1, filed under Category 3: Bulk Drug Substances Nominated Without Adequate Support. That is a record of a nomination someone filed and of what was filed with it, not an approval, not a compounding authorization and not a safety clearance; the category name is the finding. And it appears by name on FDA's list of certain bulk drug substances for use in compounding that may present significant safety risks, current to 22 April 2026, in the table of substances nominated but withdrawn. FDA's own text there reads: compounded drugs containing thymosin-alpha-1 may pose significant risk for immunogenicity for certain routes of administration and may have complexities with regard to peptide-related impurities and API characterization, and the safety-related information is inadequate for the agency to sufficiently understand the extent of any safety issues raised by the proposed compounded drug. A withdrawn nomination flagged for safety risks is the opposite of a compounding authorization. The same table lists thymosin beta-4 fragment (LKKTETQ), also known as TB-500, as a separate substance with its own separate risk statement — a first-party confirmation, in a regulatory document, that the two are not one thing.

Orphan-drug designations are real here, and a designation is not an approval. FDA's orphan designation database was searched on 1 September 2026 and returns four designations for thymalfasin, all held by SciClone Pharmaceuticals: treatment of chronic active hepatitis B, designated 3 May 1991; treatment of DiGeorge anomaly with immune defects, 8 January 1998; treatment of hepatocellular carcinoma, 6 March 2000; and treatment of stage IIb through stage IV malignant melanoma, 13 March 2006. Each of the four carries FDA's own status line, Not FDA Approved for Orphan Indication. An orphan designation is a statement that a disease is rare, plus a package of development incentives for whoever is chasing it. It is not evidence that a drug works, it is not permission to market one, and it changes nothing about the badges on this page: thirty-five years after the first of these four designations, there is still no US approval and no approved US product. One retrieval note, because this is the kind of thing that gets miscopied: the database's product-name search path returned intermittent HTTP 503 errors on some attempts, for control terms as well as for this substance, while the sponsor-name path returned results every time and the product-name path resolved on retry. Those errors were a flaky code path, not an unavailable register, and they say nothing about the substance. Searching the same database on the same day for thymosin returns a different substance, thymosin beta 4, carrying its own separate designations under its own separate sponsor — one more first-party regulatory record that the two molecules are not interchangeable.

On approvals outside the United States, this page states what was checked and stops. Thymalfasin is very widely described online as approved in thirty or more countries under a brand name. That claim was not confirmed at any primary regulator source reachable in this session. Health Canada's Drug Product Database was queried by brand name and by active ingredient and returned verified empty results for both, while control queries in the same session returned Canadian records as expected — so for Canada this is a real negative rather than a failed lookup. Other national and regional registers were attempted and did not produce a usable answer: the European Medicines Agency's search returned HTTP 401 for the search term and for its control alike, the UK and EU medicines registers returned pages whose result lists are assembled in the browser and were empty for the control as well, and the Australian register did not respond. This page therefore says neither that the peptide is approved in thirty countries nor that it is unapproved everywhere outside the United States: the US position is verified and negative, the Canadian position is verified and negative, and the rest was not established. A journal article that mentions a brand name is not a registration record, and a marketing authorization anywhere is a fact about that jurisdiction rather than a quality signal, an endorsement, a substitute for FDA review, or a route to obtaining anything.

Anti-doping. The 2026 World Anti-Doping Agency Prohibited List, in force from 1 January 2026, was retrieved on 1 September 2026 through the US Anti-Doping Agency's prohibited-list page, which links the WADA-hosted PDF. Searching that document and the accompanying 2026 Monitoring Program for thymalfasin, thymosin alpha 1, prothymosin and Zadaxin returned no hits in either. The only thymosin named anywhere on the List is thymosin-β4 and its derivatives, given with TB-500 as the example, at section S2.3 — a different molecule, and that entry says nothing about this one. Absence from the List is never a statement that a substance is permitted. The provision that reaches an unnamed substance is S0, non-approved substances, which prohibits at all times "any pharmacological substance which is not addressed by any of the subsequent sections of the List and with no current approval by any governmental regulatory health authority for human therapeutic use". Whether that description fits thymalfasin turns on the same non-US approval question this page could not settle, so the honest statement is that S0 is the provision to read and that its trigger condition is unresolved here. Anyone competing under an anti-doping code should treat the current year's List and their own anti-doping organization as the authority.

Randomized human study of this peptide is not new — a placebo-controlled hepatitis C trial was published in 1996^[1] and the largest randomized trial appeared in 2025^[17] — but none of it has produced a US approval, and every one of those trials tested the peptide in a specific diagnosed illness rather than in a healthy person. What that leaves a reader with is narrow and practical: there is no US label, no approved manufacturer, no assigned indication and no regulated quality standard behind material labelled with this compound's name.

Questions to bring to a provider

This is an unusual page to bring to a clinician, because the problem is not missing evidence — it is a body of evidence that mostly reads the wrong way for the reason people ask about it. Things worth putting on the table:

• Which molecule is actually being discussed? A conversation about "thymosin" can mean this 28-residue peptide or thymosin beta-4, and the two came out of the same 1970s extract with nothing else in common.^[8]
• The phase 3 trial in 1,106 patients with sepsis reported a hazard ratio of 0.99 and a P of 0.93 on its primary endpoint. What reading of that result would make the peptide worth considering anyway?^[17]
• Is a diagnosed condition in play at all — chronic hepatitis B, cirrhosis, sepsis — and what does the current standard of care for it offer against a randomized comparison?^[2]
• If the interest is general immune support, what measurable outcome would count as an answer, and over what period? Nothing in the published record defines one for a healthy adult.
• Does a history of autoimmune disease, active or recent infection, or malignancy change the risk picture, given that the peptide's effect on dendritic cells ran in opposite directions depending on whether the stimulus was viral or bacterial?^[11]
• Most human exposure to this peptide happened alongside another drug — entecavir, interferon, standard sepsis care. What interaction questions does that raise for the medications already in use?^[12]
• Under an anti-doping code, being unnamed on a list is not the same as being cleared. Who is the right body to ask about a substance that no section names explicitly, and what turns on the answer?

There is no shortage of published human data here. There is a shortage of published human data pointing the direction the marketing does, and a clinician who says so is reading the same trials this page does.