Thymosin Alpha-1: Sepsis Trial Results

The earlier signal

The 2013 ETASS study randomized 361 patients with severe sepsis across six hospitals in China. It compared thymosin alpha-1 plus conventional treatment with a control group and used a single-blind design. Death from any cause within 28 days was the primary outcome ^[1].

Mortality was 26.0 percent in the thymosin group and 35.0 percent in the control group. That numerical difference deserves attention, but the statistical result depended on the analysis: the nonstratified comparison had a p value of 0.062, while the survival comparison had a p value of 0.049. The relative-risk confidence interval extended from 0.54 to 1.02, crossing the no-difference value ^[1].

This is a suggestive result rather than an unambiguous demonstration. Choosing only the more favorable p value hides the uncertainty that the report itself makes visible.

The larger blinded test

TESTS, published in 2025, randomized 1,106 adults across 22 centers in China. It was double-blind and placebo-controlled. The principal question was again whether treatment reduced all-cause mortality within 28 days; the modified intention-to-treat analysis included 1,089 participants who received study treatment ^[2].

The trial did not establish a mortality advantage. It also reported no statistically significant difference in its secondary or safety outcomes. These findings are more informative about the tested sepsis population than a general description of thymosin alpha-1 as an immune modulator ^[2].

Neither study was a trial of routine immune support in healthy people. Patients receiving intensive care for sepsis cannot stand in for people seeking fewer colds, improved energy or longevity.

The correction matters

A May 2025 correction explained that updated survival information for nine participants had not been included in the original survival analyses. The corrected hazard ratio for 28-day mortality is 0.97, with a 95 percent confidence interval of 0.76 to 1.24. The original estimate was 0.99 ^[3].

The correction changes the number a careful review should cite, but it does not turn the result into a demonstrated survival benefit. The interval spans the no-difference value of one and includes effects in either direction. The correction also addresses lymphocyte data and the data-sharing statement; it is a correction, not a retraction ^[3].

Immune measurements and survival

ETASS reported greater improvement in monocyte HLA-DR, an immune-related measurement, in the treatment group ^[1]. A biological response can help explain how a treatment might work, but it cannot replace the outcome the trial was designed to test. An improved laboratory marker does not by itself mean that fewer patients die.

TESTS reported possible differences between patient subgroups ^[2]. Such signals can guide another trial, but they should not be presented as a proven selection rule. A subgroup claim needs confirmation in patients selected for that characteristic, with the intended comparison specified before outcomes are examined.

What the evidence supports

Thymosin alpha-1 has been studied in people; describing it as having no human evidence would be wrong. Describing sepsis survival benefit as established would also go beyond these trials. The useful conclusion is narrower: an earlier signal did not become a clear overall benefit in the larger blinded comparison.

The thymosin alpha-1 monograph keeps its separately dated US regulatory record. A trial result and a US supply authorization answer different questions, and neither an immune-marker change nor a foreign clinical study settles both.

Elsewhere in this index