Thymosin Alpha-1

The large TESTS sepsis trial (1,106 randomized adults at 22 Chinese centres) did not demonstrate a reduction in 28-day mortality with thymosin alpha-1: 127 of 542 treated participants (23.4%) and 132 of 547 placebo participants (24.1%) died within 28 days (corrected hazard ratio 0.97, 95% CI 0.76-1.24). Ninety-day mortality (31.0% versus 32.4%) and safety outcomes, including overall adverse and serious adverse events, also did not differ significantly between groups. PMID 40447307 is the official published correction to this same trial, not a second independent study: it revised nine participants' survival records and lymphocyte data for 35 participants after a covid-19-era data-verification disruption, and the corrected 28-day hazard ratio replaced an earlier uncorrected value of 0.99. Older uncorrected estimates should not be reused.^[1][2]

• This does not rule out every indication, but it contradicts a claim of established general sepsis survival benefit.
• The two cited source records describe one trial and its own erratum, not two independent studies; PubMed and Europe PMC index PMID 40447307 as a formal 'Erratum for' PMID 39814420 (and reciprocally 'Erratum in' on the original record), so this is not independent corroboration of the same finding.

A 2025 systematic review and meta-analysis pooled 11 randomized controlled trials of thymosin alpha-1 for sepsis (967 treated participants versus 960 controls) and found a statistically significant reduction in pooled 28-day mortality (odds ratio 0.73, 95% CI 0.59-0.90). This benefit was not statistically significant when restricted to higher-quality trials (odds ratio 0.82, 95% CI 0.65-1.03) or to multicentre trials (odds ratio 0.86, 95% CI 0.68-1.08), and the review's own trial sequential analysis concluded the current total sample size is inadequate to confirm the pooled effect. This is broadly consistent with the null result of the large TESTS trial recorded separately in this record.^[3]

• Independent full-text inspection (2026-09-09) confirms PMID 39814420 (TESTS) is one of the 11 pooled trials (Table 1 and reference list of PMC12440967 name it explicitly, matching its exact sample size and dosing regimen). TESTS supplies 56.5% of the meta-analysis's total pooled patients, 69.4% of its higher-quality subgroup, and 73.0% of its multicentre subgroup. This means the subgroup analyses that lost statistical significance are not independent corroboration of TESTS's null result from a separate body of trials; they substantially reflect TESTS's own data. The remaining four higher-quality, non-TESTS studies contribute only about 479 of the 1568 higher-quality-subgroup patients.
• A systematic review's subgroup and sensitivity findings are not a substitute for inspecting each included trial, and pooled estimates from heterogeneous small trials carry their own risk of bias.