Precision Immunotherapy to Improve Sepsis Outcomes

Giamarellos-Bourboulis. JAMA 2025. doi:10.001/jama.2025.24175

Clinical Question

• In patients with sepsis, does immunotherapy (using either anakinra or human interferon gamma) that is tailored to the type of immune dysregulation lead to a reduction in organ dysfunction at day 9?

Background

• Sepsis can involve a dysregulated immune response to an infection (SIDP, sepsis induced immune dysfunction)
• Heterogeneity in the type of dysregulation an individual patient may experience bas been demonstrated
• In particular, some patients experience a hyper-immune state whereas others demonstrate features of immunoparalysis
• Various immunotherapies have been studied in RCTs however these trials have not considered any heterogeneity of treatment effect
• Blood ferritin concentration and HLA DR expression can be used to classify the type of immune response into 3 states of immune activation:
  • 1) Macrophage activation like syndrome (MALS) in which there is excess production of interleukins with hyper-inflammation
  • 2) Sepsis-induced immunoparalysis (SIP) due to exhaustion and/or apoptosis of lymphocytes resulting in hypo-inflammation 
  • 3) Unclassified: patients who fit neither of the above

Design

• Prospective, interventional RCT phase 2b trial
• Randomisation: Allocated 1:1 into treatment or control with separate randomisation at each study site by computer generated sequencing
• Biospecimens shipped within 6 hours to a central laboratory in each country for analysis of immune state
• Blinding: The study drugs were prepared by unblinded pharmacists, all other participants (investigators, attending physicians, nursing staff and patients) were blinded to the assigned intervention
• Double Dummy: given the different formulations and administration routes of the drug involved patients in the intervention group each received a saline placebo for the drug they were not randomised too (i.e. in anakinra group, they also received a subcutaneous injection)
• Assessment: Data uploaded to central electronic record and monitored by blinded/unblinded assessors
• Primary Endpoint:
  •  Proportion of patients in each group who attained a 1.4-point or more decrease in the mean SOFA score by day 9
    • The mean SOFA change was calculated after subtracting the mean value of SOFA scores between days 2 and 9 from the baseline SOFA score on day 1
  • This was based on RCTs showing a reduction of the mean SOFA score by 1.4 points was associated with reduced 28-day mortality
• Sample size based on 40% of intervention group and 20% of control achieving the primary endpoint. With a dropout rate of 15%, 280 patients required
• If patients discharged home or lost to follow-up before day 9, the last recorded SOFA score was carried forward

Setting

•  33 study sites
  • 6 countries (Greece, Germany, Italy, Netherlands, Switzerland, Romania)
  • August 2021 to April 2024

Population

• Inclusion:
  • Adults
  • CAP, HAP, VAP, primary bacteraemia
  • Sepsis defined by Sepsis-3
  • Time from identification of sepsis to start of blind intervention <72h
  • Initial blood test confirming immune dysregulation
    • All patients had blood tests taken at enrolment – Blood tests analysed for blood ferritin and the number of HLA-DR receptors on CD45/CD14 monocytes 
      • Ferritin > 4420ng/ml classified as having MALS
      • <5000 HLA-DR molecules on CD45/CD14 monocytes and Ferritin < 4420 classified as SIP
      • Neither of the above were determined “unclassified” immune state and not eligible for the study
• Exclusion:
  • >72h from sepsis onset 
  • Sepsis from: abdominal source, UTI, meningitis, skin source, SARS-CoV pneumonia w/o MAS
  • Unclassified immune state on enrolment blood tests
  • Neutropenia
  • A past medical history that includes Primary immunodeficiency, stage IV malignancy, MS, SLE, HIV
• Participant Numbers:
  • 672 screened -> 281 included -> 135 intervention, 146 control
    • Most (n = 355) due to unclassifiable immune state (ferritin < 4420 ng/ml and HLA-DR > 5000)
• Baseline Characteristics:
• Comparing precision immunotherapy vs. placebo group:
  • Age 69 vs 70
  • Female: 32 vs 35%
  • Charlson comorbidity index > 5: 41 vs 52%
  • Immune dysregulation:
    • MALS 19 vs 16%
      • Median Ferritin: 5438 vs 6221 ng/ml
    • SIP 81 vs 84%
      • Median mHLA-DR: 3043 vs 3078 antibodies per CD45/14 cell
  • Median APACHE II: 20 vs 19 
  • Median initial SOFA score: 10 vs 9 
  • Mechanical Ventilation: 89 vs 89%
  • CRRT: 18 vs 17%
  • Time between sepsis onset and commencement of drug: 36 vs 47 hrs

Intervention

• If MALS: Anakinra IV 200mg + s/c dummy treatment (an IL-1 antagonist)
• If SIP: 100mcg s/c recombinant human interferon gamma + IV dummy treatment (immune stimulating)
  • Treatment administered daily for up to 15 days
  • If Cr Clearance < 30 mls/min then 50% of dose received

Control

• 0.9% saline used in place of drugs above

Management common to both groups

• Patients received 1 intravenous (IV) injection every 8 hours and 1 subcutaneous injection every 48 hours for 15 days
• All other care provided was best standard care

Outcome

• Primary outcome:
• Proportion of patients in each group who attained a 1.4 point or more decrease in mean SOFA score by day 9
  • 46/131 (35.1%) in precision immunotherapy group vs 26/154 (17.9%) in control (unadjusted OR 2.48; 95% CI 1.42 – 4.32, p=0.002)
  • Results consistent in model adjusted for illness severity and a sensitivity analysis whereby patients who were discharged before day 9 were excluded
  • Significant interaction demonstrated in those with higher illness severity (SOFA > 10) or Charlson Comorbidity Index > 5
• Secondary outcomes:
  • Comparing precision immunotherapy group vs placebo group
• Greater in precision immunotherapy groups:
  • Reduction in mean SOFA score of 1.4 by day 9 by each intervention group 
    • MALS: 48% vs 17.4% (p=0.04)
    • SIP: 32.1% vs 18.0% (p=0.02)
  • 1.4 point or more decrease in mean SOFA score by day 2 to 15
    • 51/131 (39.7%) vs 34/145 (23.4%) 
  • Time to reversal of SIDF also significantly lower in intervention group (subset of 125 patients)
    • Defined as sustained reduction in ferritin > 15%, or sustained increase in absolute number of HLA-DR receptors > 8000
    • Reversal of SIDF 78% vs 45% (OR 3.76, 95% CI 1.72 – 8.22)
• Less in precision group:
  • A worse infection outcome of precision group vs placebo group
  • Outcomes defined as superinfection, treatment failure or lack of resolution of infective process
    • OR 0.59 (95% CI 0.38 – 0.91) 
• No difference:
  • 28 day mortality and 90 day mortality
    • 28 day: 43.5% vs 49.7% (p=0.34)
    • 90 day: 68.7% vs 67.6% (p=0.9)
• Adverse Events:
• Significant difference in AE for intervention group vs control
  • Anaemia in Anakinra group
  • Haemorrhage in interferon gamma group

Authors’ Conclusions

• Among patients with sepsis, a precision immunotherapy strategy targeting macrophage activation–like syndrome and sepsis induced immunoparalysis improved organ dysfunction by day 9 compared with placebo

Strengths

• Multi-centre, international RCT
• Study groups well balanced at baseline
• Bias was minimised with computer-generated randomisation sequences, allocation concealment and blinding of all staff involved in care, alongside patients and assessors
• No crossover between study groups
• Low loss to follow up with 98.2% of patients included in analysis for primary outcome
• All pre-described primary and secondary outcomes were reported

Weaknesses

• The primary outcome of a SOFA score is a surrogate rather than patient-centred outcome
  • Despite achieving this end-point, there was not a statistically significant difference in 28-day mortality between the groups
• Smaller numbers (n = 48) of patients with MALS
• Of the 672 patients screened – > 50% were excluded due to having an unclassified immune state and it is unclear how this study affects this large cohort of patients 
• Limited evidence base to justify cut-offs, and dose/durations of therapies provided – would a more tailored approach to duration of therapy be better? Would other agents yield similar results?
• Not generalisable to all sepsis as analysis only included patients with pneumonia or primary bacteraemia
• The control arm consisted of only standard sepsis management, with potential heterogeneity of practice between different centres in an international trial

The Bottom Line

• This RCT demonstrated that immunotherapy targeted to the particular type of immune dysfunction in sepsis can reduce organ dysfunction by day 9. However it is not associated with a reduction in mortality at any stage
• While the results of this trial represent a promising direction for further research; the availability of HLA-DR and ferritin testing, cost and availability of interventions, and significant prevalence of patients with unclassified immune states based on current definitions all represent barriers to the wider implementation of targeted immunotherapy for sepsis at this stage

External Links

• article Precision Immunotherapy to Improve Sepsis Outcomes The ImmunoSep Randomized Clinical Trial
• editorial Precision Therapy for Sepsis The End of the Beginning?

Metadata

Summary author: Patrick McNamara
Summary date: 30th December 2025
Peer-review editor: George Walker

Picture by: Belova59/Pixabay