Thymosin Alpha 1 Sepsis Trial: What TESTS Found

The thymosin alpha 1 sepsis trial called TESTS did not show a clear reduction in deaths. About one in four patients died within 28 days in both groups: 23.4% with the peptide and 24.1% with placebo.

That matters because this was a large, carefully controlled hospital study. It tested people with sepsis, a life-threatening response to infection that damages organs. It did not test everyday immune support in healthy adults.

What did the thymosin alpha 1 sepsis trial ask?

Could adding thymosin alpha-1 to usual sepsis care reduce deaths? Thymosin alpha-1 is a peptide, a short chain of protein building blocks. Its medicine name is thymalfasin.

Researchers were interested in its effects on immune activity during severe infection. Earlier small trials had suggested benefit, but TESTS's authors described weaknesses such as small samples, unmasked treatment and differences between groups.

The question was survival, not whether a blood test changed. Our thymosin alpha-1 benefits guide covers the broader evidence across other conditions.

Who joined TESTS, and for how long?

Wu and colleagues' TESTS report, published in BMJ on January 15, 2025, describes 1,106 adults at 22 Chinese centers. Enrollment ran from September 2016 through December 2020. Eligible ages were 18–85. The median age—the age in the middle of the group—was 65.

Participants had sepsis under Sepsis-3, a definition requiring infection-related organ dysfunction. Exclusions included pregnancy, breastfeeding, organ or bone-marrow transplant, blood cancers and active autoimmune disease, where immunity attacks the body's own tissues. Recent radiotherapy, chemotherapy, medicines suppressing immunity, or higher-dose steroids also excluded patients.

Random assignment put 552 people in the treatment group and 554 in the placebo group. Placebo means a matching preparation without the active drug. Double-blind means patients and the study team did not know the assignments.

The published protocol used 1.6 mg injected under the skin every 12 hours for up to seven days. Treatment could end earlier after intensive-care discharge, death or withdrawal of consent. Both groups received usual sepsis care; follow-up continued to 90 days.

The ClinicalTrials.gov record, NCT02867267, identifies TESTS as phase 3, a larger study designed to test treatment effectiveness. This was hospital dosing under intensive monitoring, not an outpatient schedule.

What was measured, and what happened?

The primary endpoint—the main result chosen beforehand—was death from any cause within 28 days of random assignment.

The main analysis included 1,089 people who received at least one dose. Seventeen withdrew before receiving treatment. The paper calls this a modified intention-to-treat analysis.

The difference was 0.7 percentage points, with no clear statistical separation. The corrected hazard ratio, a comparison of death rates over time, was 0.97. Its 95% confidence interval, the uncertainty range around that estimate, was 0.76–1.24. It includes 1, which represents equal rates.

The corrected survival comparison had a P value of 0.82. A P value measures how compatible the data are with no treatment difference. At 90 days, deaths were 31.0% versus 32.4%, also without a clear difference. Other clinical outcomes, including new infections and days without breathing-machine support, did not differ significantly.

Why there are two PubMed records

The May 30, 2025 BMJ correction is a separate record, not another trial. Pandemic-related database problems left survival analyses using outdated information for nine patients. The corrected 28-day hazard ratio is 0.97, replacing 0.99. The correction also addresses immune-cell data errors in 35 patients. The overall conclusion remains unchanged.

What did the safety table show?

Around two in three patients in each group had an adverse event, meaning a medical problem recorded during follow-up. That does not establish that the drug caused it.

Neither comparison differed significantly: P=0.70 and P=0.38, respectively. The investigators reported no unexpected serious event related to thymosin alpha-1.

These findings describe a short course in critically ill adults. They cannot establish safety during months or years of outpatient use. The peptide overview puts these findings beside other human trials.

Did age or diabetes change the result?

Prespecified subgroups are comparisons planned before results are known. TESTS examined age, sex and six existing illnesses. These were hypertension (high blood pressure), coronary heart disease (narrowed heart arteries) and chronic obstructive pulmonary disease (lasting airflow blockage). They also included diabetes (raised blood sugar), chronic kidney disease (lasting kidney damage) and solid malignant tumour (cancer outside blood or marrow).

Some headline estimates pointed in different directions:

The younger group leaned toward more deaths with treatment; the diabetes group leaned toward fewer. Neither supports a firm treatment rule.

The interaction tests, which ask whether treatment effects differ between subgroups, gave P=0.01 for age and P=0.04 for diabetes. Neither met the paper's stricter P=0.006 threshold for multiple comparisons. These findings are hypothesis-generating: ideas for another trial, not proof of benefit or harm in a particular age group.

What did critics and later research add?

Powers and Natanson's accompanying BMJ editorial emphasized enrolling the right number and type of patient. TESTS's authors acknowledged that their study was designed around an ambitious eight-percentage-point mortality difference. Smaller effects remain uncertain, and treatment timing was difficult to establish.

As of October 1, 2026, the day this article was written, later evidence includes a review published May 12, 2026. It compared several immune treatments across 76 trials involving 22,194 patients, not all receiving thymosin alpha-1. Thymosin alpha-1 alone did not show a statistically significant mortality benefit in that broader analysis.

As of October 1, 2026, the day this article was written, an August 5, 2026 BMC Infectious Diseases study adds a narrower finding. It reviewed records of 327 intensive-care patients with tuberculosis-related sepsis, caused by the bacterial infection tuberculosis. Treatment was associated with lower mortality, but patients were not randomly assigned. Differences in illness and treatment selection could explain the association; it does not overturn TESTS or establish causation.

What does TESTS mean for a compounded prescription?

TESTS gives a useful answer about one hospital protocol. It does not establish routine immune benefits or show that a compounded preparation performs like the study medication.

As of October 1, 2026, the day this article was written, thymosin alpha-1 has no FDA-approved U.S. product. FDA's evaluation dated November 15, 2024 documents that status; its hepatitis B orphan-drug record still lists no approval. The separate U.S. regulatory record explains the history.

Thymosin alpha-1 prescribed through Promise is dispensed by a licensed U.S. compounding pharmacy as a compounded medication, meaning the pharmacy prepares it for an individual prescription. It is different from an FDA-approved product: the formulation offered here is not FDA-approved. A licensed provider may still prescribe a compounded formulation; that decision is between the patient and the doctor.

What belongs in the provider conversation?

A useful conversation connects the intended goal with the studied population, transplant or autoimmune history, current medicines and a plan for follow-up. The prescriber sets any individual dose and duration.

At Promise, a licensed U.S. provider reviews every request, and not everyone qualifies. The evidence question remains specific: does this hospital study actually answer the question being asked for this person?