Thymosin Alpha-1 (Thymalfasin)

Identity and origin

Thymosin alpha-1 is a chain of 28 amino acids with an acetyl group on the N-terminal serine. PubChem records it as CID 16130571 with the formula C129H215N33O55 and a molecular weight of 3108.3 g/mol; the FDA Global Substance Registration System lists the same molecule under UNII W0B22ISQ1C with CAS registry number 62304-98-7 and the sequence SDAAVDTSSEITTKDLKEKKEVVEEAEN. Those two records agree, and the formula reconciles with the acetylated sequence rather than the bare peptide, which would be C127H213N33O54. The synthetic form carries the international non-proprietary name thymalfasin.

Isolation came first, from thymosin fraction 5 of calf thymus in 1977. The peptide is endogenous: it arises from proteolytic cleavage of prothymosin alpha in a range of mammalian tissues, which is the point Romani and colleagues make in their 2007 review for the New York Academy of Sciences. In healthy volunteers the circulating concentration sits below the quantification limit of the standard ELISA, reported as 0.10 micrograms per litre by Rost and colleagues in 1999, so baseline serum measurement is not a straightforward matter.

No USP drug substance monograph exists for the substance, and the FDA review of December 2024 records that it is not recognised in the European or Japanese Pharmacopoeias either. That same review notes solubility in water of up to 2 mg/mL for the free base and 1 mg/mL for the acetate salt, and describes both forms as not well characterised, citing missing data on impurities, aggregates and endotoxin levels.

Where it is and is not a medicine

Regulatory position is the fact most often stated incorrectly. The FDA states plainly that thymosin alpha-1 is not approved in the United States, Japan or Europe, with Italy the exception. Beyond that, the agency reports what a manufacturer has claimed rather than what it can confirm: the 2014 annual report of the company that markets the peptide describes approval across parts of the Asia-Pacific region, Latin America, Eastern Europe and the Middle East, and the FDA states it was unable to independently verify those claims of approval in all the specified countries.

Compounded preparations are a separate matter from approval. The FDA review found no evidence of compounded products containing the substance in the published literature or in outsourcing facility reports, but did find them marketed online in the United States as injectables and nasal sprays, offered for hepatitis B, hepatitis C, HIV, chronic fatigue, inflammation, sepsis, COVID-19, Lyme disease, allergies, cancer, asthma, COPD and psoriatic arthritis. One such website described its product as FDA-approved. The agency's response in the briefing document is a single sentence: no drug products containing thymosin alpha-1 have been approved.

Sixty-five studies naming thymosin alpha-1 or thymalfasin are registered on ClinicalTrials.gov. The largest currently recruiting is a phase 3 adjuvant trial in high-risk stage II and III colorectal cancer with a planned enrolment of 2,500 and estimated primary completion in March 2027. A United States phase 2 trial in COVID-19 patients with lymphocytopenia was terminated after enrolling 56, the sponsor recording slow enrolment and the departure of the principal investigator. Most registered activity is oncology adjunct work conducted in China.

Pharmacokinetics as measured

Human kinetics rest on one small crossover study. Rost and colleagues gave nine Caucasian volunteers 900 micrograms per square metre subcutaneously in a randomised three-way crossover comparing three formulations then on the market, and dosed for five days. Time to maximum concentration fell between one and two hours, peak concentrations between 30 and 80 micrograms per litre, and elimination half-life under three hours. Area under the curve and peak concentration did not differ between single and multiple dosing, indicating no accumulation over five days. Apparent volume of distribution of 30 to 40 litres is consistent with distribution confined to extracellular fluid.

Formulation mattered in that study, a detail usually lost when the half-life figure is quoted alone: area under the curve and peak concentration were similar for two of the three preparations and higher for the third. The rodent numbers sit in a different range entirely. The FDA review cites work by Wang in 2018 and Shen in 2019 reporting a half-life of 1.9 to 3 minutes in rats given roughly 0.175 mg/kg intravenously, and states that the subcutaneous nonclinical profile is unknown. Species and route differ, so the two figures are not in conflict; they are not comparable either.

Sepsis: the largest randomised dataset

Two trials dominate. The ETASS trial, published by Wu and colleagues in Critical Care in 2013, randomised 361 patients with severe sepsis across six Chinese teaching hospitals to subcutaneous thymosin alpha-1 or control, single-blind. Twenty-eight-day all-cause mortality was 26.0% in the treated arm against 35.0% in the control arm, a relative risk of 0.74 with a confidence interval of 0.54 to 1.02 and a p value that crossed the conventional threshold on one test and not the other, 0.049 by log rank and 0.062 unstratified. Monocyte HLA-DR expression improved more in the treated arm at days three and seven.

Nine years later the same principal investigator ran the confirmatory trial. TESTS, published in the BMJ in January 2025, randomised 1,106 adults meeting sepsis-3 criteria at 22 centres to subcutaneous thymosin alpha-1 or placebo every twelve hours for seven days, double blinded. Twenty-eight-day all-cause mortality was 23.4% against 24.1%, a hazard ratio of 0.99 with a confidence interval of 0.77 to 1.27 and a p value of 0.93. No secondary or safety outcome differed significantly. The authors state the trial found no clear evidence that the drug decreases 28-day mortality in adults with sepsis.

Prespecified subgroup analysis in TESTS produced two interactions worth recording without over-reading: participants under 60 had a hazard ratio of 1.67, those 60 and over 0.81, interaction p value 0.01; participants with diabetes had a hazard ratio of 0.58 against 1.16 without, interaction p value 0.04. A 2025 meta-analysis by Gu and colleagues pooled 11 randomised trials covering 967 treated and 960 control patients and reported an odds ratio of 0.73 for 28-day mortality overall, but 0.86 restricted to multicentre trials and 0.82 restricted to high-quality trials, neither reaching significance, with trial sequential analysis indicating the accumulated sample remains inadequate. The company that markets the peptide funded TESTS, and several of its authors declare grants from it.

Hepatitis B: the oldest claim and the least settled

Chien and colleagues randomised 98 patients with chronic hepatitis B in Taiwan to 26 weeks of treatment, 52 weeks of treatment, or observation, and reported in Hepatology in 1998 that complete virological response assessed 18 months after entry was 40.6% in the 26-week arm, 26.5% in the 52-week arm and 9.4% in the untreated arm. Response rates were similar across the three groups at the end of therapy and diverged afterward. That 40.6% against 9.4% comparison is the single number that circulates most widely about this compound, and it comes from an unblinded trial with an untreated rather than placebo control.

A double-blind placebo-controlled trial reported the following year did not reproduce it. Mutchnick and colleagues randomised 97 patients with HBeAg-positive chronic hepatitis B to 1.6 mg twice weekly or placebo for six months with six months of follow-up, and found complete response in 7 of 49 treated patients (14%) against 2 of 48 on placebo (4%), p value 0.084. Their own summary states the results do not confirm observations of treatment efficacy reported in other clinical studies. The FDA review counts five monotherapy studies from 1998 to 2005 and describes the results as mixed.

Combination work has been no more decisive. Two studies comparing pegylated interferon with and without the peptide over 48 weeks found virologic response rates of 35% against 40% in one and 40% against 42.3% in the other, neither statistically separated. A randomised open-label trial of entecavir with and without the peptide over 52 weeks found undetectable HBV DNA in 64.6% against 69.6% at week 52 and 87.7% against 85.5% at week 104. A 2020 meta-analysis by Peng and colleagues covering seven trials and 1,144 subjects with HBV-related cirrhosis reported a higher complete response with combination therapy, relative risk 1.18 with a confidence interval of 1.07 to 1.30, while finding no separation in HBV DNA undetectability at 52 weeks. In a 2022 randomised open-label trial in 120 patients with HBV-related acute-on-chronic liver failure, 90-day transplantation-free survival was 75.0% against 53.4%, though the authors report no significant difference under competing-risk analysis.

COVID-19: two cohorts pointing opposite ways

Liu and colleagues reviewed 76 patients with severe COVID-19 retrospectively and reported in Clinical Infectious Diseases in 2020 that mortality was 11.11% among treated patients against 30.00% among untreated, p value 0.044, alongside falls in PD-1 and Tim-3 expression on CD8 T cells and rises in T-cell receptor excision circles. The paper is retrospective, and its title asserts a causal reduction in mortality that its design cannot establish.

A larger cohort published the following year in Frontiers in Immunology ran the other way. Liu and colleagues, a different group, examined 2,282 patients across five tertiary hospitals in Hubei and found every crude outcome worse in the treated group, including in-hospital mortality of 62 of 306 against 271 of 1,976. After adjustment for confounders, treatment was associated with a higher non-recovery rate, odds ratio 1.5 with a confidence interval of 1.1 to 2.1. Confounding by indication is the obvious reading of both studies in opposite directions, and neither is randomised.

Meta-analyses have not resolved it. The FDA review tabulates four, of which three found no decrease in mortality and one found an association with lower mortality but no reduction in mechanical ventilation or length of stay. Treatment guidelines from the Infectious Diseases Society of America, the National Institutes of Health and the World Health Organization do not discuss the compound for COVID-19 at all. A registered prospective trial in 194 adults with end-stage renal disease evaluating prevention had not reported efficacy results at the time of the FDA review.

What the 2024 FDA review concluded

Twelve proposed uses were assessed: hepatitis B, hepatitis C, HIV, COVID-19, depressed vaccine response, malignant melanoma, hepatocellular carcinoma, non-small-cell lung cancer, sepsis, infections after haematopoietic stem cell transplantation, COPD, and myalgic encephalomyelitis and chronic fatigue syndrome. For the last of these the agency identified no clinical studies at all. For each of the other eleven the conclusion was some form of insufficient evidence, with the recurring criticisms being small sample sizes, absent or outdated comparators, unblinded designs, and concomitant therapies that made the contribution of the peptide impossible to isolate.

Safety received a separate treatment. Across clinical studies at 1 to 16 mg subcutaneously for up to twelve months, the agency records no significant adverse events attributable to the drug, with local irritation and injection-site discomfort most common. It also lists specific signals: alanine aminotransferase flares in chronic hepatitis B, thyroid-stimulating hormone abnormalities in chronic hepatitis C, and fatal immune haemolytic anaemia and engraftment failure in a small transplant series reported by Perruccio in 2010. Immunogenicity is flagged as an unquantified risk for a 28-residue peptide given subcutaneously, amplified by aggregation and synthesis-related impurities.

On 4 December 2024 the Pharmacy Compounding Advisory Committee voted on whether the free base and acetate forms should be placed on the 503A bulks list. The tally recorded by the designated federal officer was four yes, seventeen no, zero abstentions. Committee members who voted no cited inadequacy of the data relative to the FDA presentation and, in one case, the sourcing of bulk substance from outside the United States. Members who voted yes cited access for rare indications. The vote is advisory, and the compounding status of the substance has continued to move since.

What is not known

The published record does not establish that thymosin alpha-1 changes mortality or any other hard clinical outcome in any population. The one adequately powered double-blind trial, TESTS, was null on its primary endpoint at 1,089 patients, and the sepsis meta-analysis that reports a pooled benefit loses it when restricted to multicentre or high-quality trials. The hepatitis B literature is old, largely unblinded, and used HBV DNA assays far less sensitive than current ones, so those response rates cannot be read across to contemporary practice; the two placebo-controlled comparisons available are negative or non-significant. No published randomised trial of the compound in COVID-19 was located, only cohorts pointing in opposite directions. Human pharmacokinetics rest on nine volunteers in a single 1999 study, and no subcutaneous animal pharmacokinetic profile exists at all. Immunogenicity has not been characterised. Long-term exposure beyond twelve months is undocumented, as is use in children, in pregnancy or lactation, and in anyone undergoing deliberate immunosuppression, where the FDA identifies a specific concern about graft-versus-host disease and engraftment failure. The trial base is also geographically narrow: the sepsis evidence is entirely Chinese, and the FDA notes that neither the patient population nor the concomitant therapies used may transfer. Thymosin alpha-1 is not an approved medicine in the United States, and material sold outside a registered trial has not been subject to the identity, purity, aggregation or sterility controls that apply to investigational supply.

Questions