Thymosin Alpha-1 for Sepsis: Critical Care Data
Thymosin Alpha-1
23.4% vs 24.1%
28-day mortality in the largest thymosin alpha-1 sepsis trial showed no overall difference from placebo in 1,089 ICU patients.
Wu et al., BMJ, 2025
Wu et al., BMJ, 2025
View as imageSepsis kills approximately 11 million people per year worldwide, and despite decades of ICU advances, no immunomodulatory drug has become standard treatment. Thymosin alpha-1, a 28-amino-acid peptide originally isolated from the thymus, emerged as one of the most studied candidates for reversing the immune paralysis that drives late-stage sepsis deaths. Early randomized controlled trials (RCTs) and meta-analyses reported mortality reductions of 26-41%. Then, in January 2025, the largest and most rigorous trial to date found no overall survival benefit. This article traces the full arc of that evidence, from the biological rationale to the phase 3 results that reshaped the field.
Key Takeaways
- The TESTS phase 3 trial (1,089 patients, 22 centers) found 28-day mortality of 23.4% with thymosin alpha-1 versus 24.1% with placebo (HR 0.99, P=0.93)[1]
- Earlier meta-analyses of smaller trials reported pooled risk ratios of 0.59-0.68 favoring thymosin alpha-1[2][3]
- The ETASS trial (361 patients, 2013) showed a trend toward lower mortality (26.0% vs 35.0%, P=0.062) and increased monocyte HLA-DR expression[4]
- Subgroup analyses suggest thymosin alpha-1 may benefit patients over 60 (HR 0.81) and those with diabetes (HR 0.58), but these are exploratory findings[1]
- A 2025 meta-analysis of 11 RCTs (1,927 patients) found an overall OR of 0.73 for 28-day mortality, but this signal disappeared in high-quality trial subgroups[5]
- The central question has shifted from "does thymosin alpha-1 work in all sepsis patients?" to "which sepsis patients might benefit?"
Why Sepsis Researchers Turned to Thymosin Alpha-1
Sepsis is not a single disease. It begins as an overwhelming inflammatory response to infection, but in many patients the immune system swings in the opposite direction within days, entering a state called immunoparalysis. This second phase is characterized by reduced monocyte HLA-DR expression, T-cell exhaustion, and increased susceptibility to secondary infections. Most late sepsis deaths occur during this immunosuppressed window.[6]
Thymosin alpha-1 attracted attention because its mechanism targets exactly this problem. The peptide activates dendritic cells through Toll-like receptor 9 (TLR9) signaling, triggers the p38/NF-kB pathway to promote interleukin-12 production, and upregulates HLA-DR on monocytes.[7][8] It also activates tryptophan catabolism via indoleamine 2,3-dioxygenase (IDO), establishing a balance between inflammation and tolerance rather than simply boosting one side.[9] This dual action made it theoretically suited for a disease where the immune system is simultaneously dysregulated in both directions.
For a detailed breakdown of how thymosin alpha-1 matures T-cells, see our dedicated article on the mechanism.
The ETASS Trial: The First Large RCT
The Efficacy of Thymosin Alpha 1 for Severe Sepsis (ETASS) trial, published in Critical Care in 2013, enrolled 361 patients across six tertiary hospitals in China between 2008 and 2010.[4] Patients received subcutaneous thymosin alpha-1 (1.6 mg twice daily for seven days) or standard care alone.
The primary outcome was 28-day all-cause mortality. In the thymosin alpha-1 group, 26.0% of patients died compared to 35.0% in the control group. The relative risk was 0.74 (95% CI 0.54 to 1.02), which crossed the threshold of statistical significance in the log-rank test (P=0.049) but not in the nonstratified analysis (P=0.062). This borderline result generated both enthusiasm and skepticism.
Where ETASS provided clearer data was on immune markers. Monocyte HLA-DR expression improved faster in the thymosin alpha-1 group, with a mean difference of 3.9% on day 3 (95% CI 0.2 to 7.6%, P=0.037) and 5.8% on day 7 (95% CI 1.0 to 10.5%, P=0.017). This aligned with the biological rationale: the peptide appeared to reverse immunoparalysis at the cellular level, even if the survival benefit remained uncertain.
ETASS had limitations. It was single-blind (only patients were blinded), conducted exclusively in Chinese ICUs, and the control group received no placebo injection. These design features left room for bias.
Meta-Analyses: What Pooled Data Showed
Between 2015 and 2025, three major meta-analyses pooled the available RCT data on thymosin alpha-1 for sepsis. Their conclusions evolved as larger trials entered the evidence base.
Li et al. (2015) pooled 12 controlled trials involving 1,480 patients and found a risk ratio for all-cause mortality of 0.68 (95% CI 0.59-0.78, P<0.00001).[2] The signal was strong, but the authors noted that most included trials were small, single-center, and of low methodological quality.
Liu et al. (2016) expanded the search to 19 RCTs across multiple databases, including Chinese-language literature. Among 10 RCTs reporting mortality (530 patients), the pooled risk ratio was 0.59 (95% CI 0.45-0.77, P=0.0001).[3] This analysis also found that thymosin alpha-1 increased HLA-DR expression (SMD 1.23, 95% CI 0.28-2.18, P=0.01) based on 8 studies of 721 patients, and beneficially shifted CD3+ and CD4+ T-cell counts. The twice-daily dosing schedule did not improve APACHE II scores, while once-daily dosing did. This suggested dosing frequency might matter, though the comparison was across trials rather than within a single trial.
Gu et al. (2025) incorporated the TESTS trial data and analyzed 11 RCTs with 1,927 total patients. The overall odds ratio for 28-day mortality was 0.73 (95% CI 0.59-0.90, P=0.003).[5] But when restricted to high-quality RCTs, the benefit disappeared (OR 0.82, 95% CI 0.65-1.03, P=0.09). Multi-center trials also showed no significant effect (OR 0.86, 95% CI 0.68-1.08, P=0.20). Trial sequential analysis confirmed the total sample size remained inadequate for a definitive conclusion.
The pattern across these analyses is consistent: the mortality signal that appeared robust in small, lower-quality trials attenuated as larger, better-designed studies entered the pool.
The TESTS Trial: Phase 3 Results
The Efficacy and Safety of Thymosin Alpha-1 for Sepsis (TESTS) trial, published in the BMJ in January 2025, was designed to resolve the ambiguity left by ETASS and the meta-analyses.[1] It enrolled 1,106 adults with sepsis (by Sepsis-3 criteria) across 22 centers in China between September 2016 and December 2020. Unlike ETASS, TESTS was double-blinded and placebo-controlled.
Patients received subcutaneous thymosin alpha-1 (1.6 mg) or matching placebo every 12 hours for seven days. The primary outcome was 28-day all-cause mortality.
The results were unequivocal for the overall population: 127 of 542 patients (23.4%) died in the thymosin alpha-1 group versus 132 of 547 (24.1%) in the placebo group. The hazard ratio was 0.99 (95% CI 0.77-1.27, P=0.93). No secondary outcome, including 90-day mortality, ICU-free days, or organ failure scores, differed between groups.
Two prespecified subgroup analyses produced the most discussed findings. Age showed a significant interaction (P=0.01): patients under 60 had a hazard ratio of 1.67 (95% CI 1.04-2.67), suggesting possible harm, while patients 60 and older had a hazard ratio of 0.81 (95% CI 0.61-1.09). Diabetes status also showed interaction (P=0.04): patients with diabetes had a hazard ratio of 0.58 (95% CI 0.35-0.99) while those without diabetes showed no benefit (HR 1.16, 95% CI 0.87-1.53).
These subgroup signals must be interpreted with caution. They were prespecified, which gives them more credibility than post-hoc analyses, but they were not powered for individual confirmation. The apparent harm in younger patients and benefit in diabetic patients could reflect genuine biological heterogeneity or statistical noise in smaller subgroups.
The Subgroup Question: Who Might Benefit?
The shift from "does it work?" to "who does it work for?" represents the current frontier of thymosin alpha-1 sepsis research. The Gu et al. 2025 meta-analysis conducted heterogeneity of treatment effects (HTE) analysis using individual patient data from two large RCTs (representing 75% of total patients). Potential benefits appeared in patients with cancer (moderate credibility per ICEMAN assessment), diabetes (low credibility), and coronary heart disease (low credibility).[5]
The diabetes finding aligns with a biological hypothesis. Diabetes impairs multiple arms of innate immunity, including monocyte function and cytokine production. These patients may enter sepsis with a pre-existing immune deficit that thymosin alpha-1 is better positioned to reverse. The cancer subgroup finding makes similar mechanistic sense: tumor-bearing patients often have suppressed cellular immunity.
Pei et al. (2018) argued that the central problem with sepsis immunotherapy trials is patient selection.[6] Sepsis is so heterogeneous that treating all septic patients identically obscures treatment effects that may be real in specific subpopulations. Their recommendation was to enrich future trials for patients with documented immunosuppression, measured by low monocyte HLA-DR or lymphopenia, rather than enrolling all-comers.
This enrichment strategy has not yet been tested in a large RCT. Until it is, the subgroup signals remain hypothesis-generating.
How Thymosin Alpha-1 Affects Immune Markers in Sepsis
Separate from the mortality question, the immunological data from sepsis trials is more consistent. Across multiple studies, thymosin alpha-1 treatment increased monocyte HLA-DR expression, a marker that predicts susceptibility to secondary infections and mortality when low.[3][4]
The Liu 2016 meta-analysis quantified this: across 8 studies of 721 patients, thymosin alpha-1 increased HLA-DR with a standardized mean difference of 1.23 (95% CI 0.28-2.18).[3] CD3+ and CD4+ T-cell counts also improved. Pro-inflammatory cytokines IL-6 and TNF-alpha decreased, while immunoregulatory IL-10 shifted.
The mechanistic picture is supported by preclinical work. Romani et al. demonstrated that thymosin alpha-1 activates dendritic cells through TLR9, promoting Th1 immune responses and interleukin-12 production.[7] In follow-up work, they showed it also induces IDO-dependent tryptophan catabolism, creating a regulatory environment that balances inflammation and tolerance.[9] Naylor et al. detailed how thymosin alpha-1 synergizes with cytokines including interferons and IL-2, amplifying immune restoration without triggering the cytokine storm that complicates other immunostimulatory approaches.[8]
The gap between these immune marker improvements and mortality outcomes is one of the key puzzles. The peptide appears to do what it is supposed to do at the cellular level, but translating that into survival benefit in unselected sepsis populations has proven elusive.
Thymosin Alpha-1 Beyond Sepsis: Related Critical Care Applications
Thymosin alpha-1's immune-modulating properties have been tested in adjacent critical care settings. Tian et al. (2025) published a meta-analysis of thymosin alpha-1 in severe acute pancreatitis (SAP), finding that it reduced inflammatory markers and prevented secondary infections in these critically ill patients.[10] SAP shares immunological features with sepsis, including systemic inflammation followed by immunosuppression, making these results mechanistically relevant.
During the COVID-19 pandemic, thymosin alpha-1 was used in severely ill patients, particularly in China. Matteucci et al. (2021) showed that the peptide mitigated cytokine storm responses in blood cells from COVID-19 patients in vitro, reducing TNF-alpha and IL-6 production while preserving interferon-gamma signaling.[11] Minutolo et al. (2023) found that thymosin alpha-1 restored immune homeostasis in lymphocytes during post-acute COVID-19, rebalancing CD4+/CD8+ ratios and reducing markers of T-cell exhaustion.[12]
These findings reinforce the peptide's core mechanism, restoring dysregulated immunity, while highlighting that sepsis may not be the only critical illness where this mechanism is relevant. For context on how peptide biomarkers are used to detect sepsis in the first place, see our article on procalcitonin as a sepsis biomarker. For thymosin alpha-1's role in recovering from viral infections specifically, see thymosin alpha-1 for post-viral immune recovery.
Dosing, Safety, and Practical Considerations
Across sepsis trials, the standard protocol has been subcutaneous thymosin alpha-1 at 1.6 mg (the dose used in the commercial product Zadaxin) given either once or twice daily for five to seven days. The TESTS trial used twice-daily dosing for seven days.[1]
Safety data across trials is reassuring. The ETASS trial recorded no serious drug-related adverse events.[4] The TESTS trial found no statistically significant differences in adverse events between thymosin alpha-1 and placebo groups.[1] Dominari et al. (2020) noted in their comprehensive review that the peptide has a favorable safety profile across more than 30 years of clinical use in various conditions.[13]
The Liu 2016 meta-analysis found an interesting dosing signal: once-daily thymosin alpha-1 significantly reduced APACHE II scores (SMD -0.80, 95% CI -1.14 to -0.47, P<0.0001), while twice-daily dosing did not (SMD 0.30, 95% CI -0.10 to 0.70, P=0.14).[3] This counter-intuitive finding has not been explored in head-to-head dosing trials and could reflect confounding by trial quality rather than a true dose-response relationship.
Thymosin alpha-1 is approved in over 35 countries for hepatitis B treatment and as a vaccine adjuvant, but it is not approved anywhere specifically for sepsis. Its use in this context remains off-label, driven by the earlier positive signal from smaller trials. The research into thymosin alpha-1 in cancer immunotherapy is also relevant, given the overlap between cancer-related immunosuppression and the immune dysfunction seen in sepsis.
What Comes Next
The TESTS trial did not close the book on thymosin alpha-1 for sepsis. It closed the book on treating all sepsis patients identically with the peptide. The subgroup signals in older patients and those with diabetes, combined with the consistent HLA-DR data and the Gu et al. meta-analysis identifying cancer patients as potential responders, point toward a precision medicine approach.
Mao et al. (2023) argued that thymosin alpha-1 should be "reimagined" for the immuno-oncology era, where patient selection based on immune biomarkers is standard practice.[14] Applying this same biomarker-driven strategy to sepsis, specifically enriching trials for patients with documented immunosuppression (low mHLA-DR, lymphopenia, or elevated immunosuppressive markers), could reveal whether the peptide has a role in a targeted population that broader trials missed.
No such enrichment trial has been registered as of March 2025. The gap between the biological plausibility, the immune marker data, and the phase 3 mortality result remains the defining tension in this field.
The Bottom Line
Thymosin alpha-1 for sepsis went through a complete evidence arc: promising mechanism, encouraging small trials, positive meta-analyses, and then a definitive phase 3 trial showing no overall mortality benefit. The biological signal (improved HLA-DR, T-cell recovery) is real and reproducible. The clinical translation to unselected sepsis populations is not. The open question is whether biomarker-selected subpopulations, particularly older patients, those with diabetes, or those with cancer, represent genuine responders. That question has not been answered by a prospective trial.