Targeted precision therapy for sepsis! Is it here?
ImmunoSep trial shows impressive results ... until day 90 anyway
“Sepsis” is a circularly defined syndromic concept that clinicians have historically used to express their belief that the cause of a person’s severe illness is an infection.
Within that simplistic label hides enormous diversity in the spectrum of bacteria (between and within species) that can produce severe infections, and in the immune and physiologic responses that infected people mount against them.
Even the core concept of “immune dysregulation” in sepsis obscures the complexity of multiple pathways of immune overactivation and immunosuppression occurring in the same patient simultaneously.
Much of the past 30+ years of sepsis research has glossed over this complexity. Rather, compounds have been identified that act on specific pathways to mitigate sepsis in animal models (anti-TNF-α; anti-IL-1; monoclonal antibodies; recombinant activated protein C; anti-endotoxin; etc., etc). The compounds are then tested in heterogeneous groups of human beings with “sepsis”, dooming the trials to failure: even if certain patients benefited, their signal is lost in the noise of nonresponders.
No such targeted intervention has yet been shown to be effective in an undifferentiated group of patients with “sepsis.”
The New School of Sepsis Research
The diversity of sepsis has been increasingly analyzed and mapped at the level of RNA transcription and protein expression, referred to as RNA transcriptomics or proteomics.
This process itself yields massive outputs of complex data for each patient. These can be reductively agglomerated into so-called “endotypes,” driven largely by rational but ultimately arbitrary decisions by the relevant analysts, based on biological plausibility.
Iterating on previously observed phenotypes of sepsis, one current rubric synthesizes the expanding universe of biodata into three simplified patterns of immune dysfunction:
Hyperinflammation/immune activation: including the activated macrophage syndrome, with excess IL-1 production, very high ferritin, often fevers, coagulopathy, and shock, in some ways resembling the blood disorder HLH;
Induced immunoparalysis: in which anti-inflammatory programs are suspected to operate excessively; lymphocytes act weakly and die prematurely;
Mixed or “other” states not meeting either of these patterns.
Like the underlying transcription and expression of proteins, these processes exist on a spectrum, and patients can transition between them or manifest aspects of each simultaneously.
(Or so the story goes: the substrate here consists of fantastically huge aggregated datasets of individual patients’ gene expression, and there is some inevitable curation, lensmaking, and storytelling that investigators must perform to make any sense at all from it.)
A major driver for the traditional (failed) “all comers” approach to sepsis research was the relative ease of enrolling undifferentiated patients into randomized trials. If they met established clinical criteria (e.g., Sepsis-3), they were eligible.
Clearer elucidation of illness profiles (endotypes) in sepsis opened the door to therapies targeted to individual septic patients’ immunologic derangements, but identifying them requires more effort and expense than the traditional approach.
The ImmunoSep Trial
At 33 sites in six European nations, 281 patients meeting Sepsis-3 criteria were enrolled; 276 were randomized to receive either:
Anakinra (IV) (or placebo) if they had ferritin levels >4,420 ng/mL (associated with the macrophage activation endotype);
Subcutaneous recombinant human interferon gamma (or placebo), if they had ferritin ≤4,420 ng/mL and fewer than 5,000 HLA-DR molecules (associated with the immunoparalysis type).
Patients not meeting either criterion were not enrolled.
Because of the differing preparations (subcu vs IV), both drugs had a placebo, and each patient got either a drug and a placebo of the other, or two placebos (a double-dummy design).
Results
After 9 days, more patients in the anakinra group had experienced a decrease in SOFA score by ≥1.4 points (35% vs 18%).
Earlier on (days 2-15), the drop was even more pronounced (39.7% vs 23.4%).
Mortality at 28 days was nonsignificantly lower in the anakinra group (43.5% vs 49.7%), but not at 90 days (68.7% vs. 67.6%).
Investigators assessed more patients receiving anakinra as having cleared their infections by day 15 (44% vs. 32%).
The site investigators made all these determinations (i.e., unblinded outcome assessments), but were supposed to be blind to treatment assignments (i.e., this was a double-blind, double-dummy design).
Anakinra for Sepsis + High Ferritin?
IL-1 is produced by activated macrophages. IL-1 binds to target cells (e.g., macrophages, leukocytes, endothelial cells) and triggers transcription and amplification of pro-inflammatory mediators like TNF, IL-6, IL-12, and chemokines.
(Activated macrophages also produce ferritin in high quantities and stimulate other cells to do so, and indirectly increase ferritin through increased hemophagocytosis and cell death.)
Anakinra is a recombinant IL-1 receptor antagonist; it blocks the binding of IL-1 after it is produced by activated macrophages, attenuating these pro-inflammatory amplification pathways.
Anakinra failed to show a survival benefit in two large phase 3 randomized trials conducted in the 1990s enrolling almost 1,600 patients with sepsis [1, 2], one of which was stopped for futility.
However (the present authors point out), a 2016 post hoc analysis of those trials found that among patients with hepatobiliary dysfunction and disseminated intravascular coagulation as criteria for macrophage activation–like syndrome,
Patients treated with anakinra had a significantly lower 28-day mortality than patients treated with placebo (35% vs 65%).
Conclusions
Analysis of massive datasets of patient-level RNA transcription and protein expression during episodes of clinically-defined sepsis will provide the grist for a generation of academics toiling in the research mills.
Based on the findings of the ImmunoSep trial, it could lead to a better understanding of the heterogeneity of sepsis, and even (dare we say it?) effective targeted treatments, tailored to individual patients.
No new phase 3 trial of anakinra as a targeted sepsis therapy in adults is currently listed on clinicaltrials.gov.
When we see a decline in 90-day mortality, not just the organ failure scores, we will definitely break out the exclamation points.
References
Precision Immunotherapy to Improve Sepsis Outcomes: The ImmunoSep Randomized Clinical Trial. JAMA 2025
Angus DC. Precision Therapy for Sepsis. JAMA. Published online December 8, 2025. doi:https://doi.org/10.1001/jama.2025.24099
Fisher CJ. Recombinant Human Interleukin 1 Receptor Antagonist in the Treatment of Patients With Sepsis Syndrome. JAMA. 1994;271(23):1836. doi:https://doi.org/10.1001/jama.1994.03510470040032
Post hoc Reanalysis of anakinra trials: https://journals.lww.com/ccmjournal/abstract/2016/02000/interleukin_1_receptor_blockade_is_associated_with.5.aspx
Cavaillon J, Singer M, Skirecki T. Sepsis therapies: learning from 30 years of failure of translational research to propose new leads. EMBO Molecular Medicine. 2020;12(4). doi:https://doi.org/10.15252/emmm.201810128