Skip to content
PaperFren

Critical care

Genetic IL-6 receptor blockade links to lower sepsis risk

Hamilton FW, Thomas M, Arnold D, et al. · PLoS medicine · 2023

Open access · cc by · source: Europe PMC

Mendelian randomisation treating IL6R variants as a natural experiment for receptor blockade found lower odds of sepsis (OR 0.80) and critical-care sepsis (OR 0.48) in UK Biobank, with a similar-sized protective association for severe COVID-19.

Study at a glance

Design
Mendelian randomisation — IL6R genetic instruments as proxies for lifelong IL-6 receptor blockade; UK Biobank primary
N
N=486484 · UK Biobank ≈486,484 adults including 11,643 sepsis cases; secondary FinnGen and COVID-19 HGI
Population
UK Biobank adults (with secondary European/COVID genetics cohorts)
Outcome
Odds of sepsis (and severe sepsis / severe COVID-19 phenotypes)

Structured fields used in claim comparison tables when every cited study has a complete layer.

Key findings

Genetically proxied IL6R blockade associated with lower odds of sepsis (OR 0.80, 95% CI 0.66–0.96) and stronger associations for more severe phenotypes such as critical-care admission with sepsis (OR 0.48). A protective association of similar magnitude was confirmed for severe COVID-19 (OR 0.69), where IL-6 blockade is already recommended in critical illness.

Methodology

Authors used Mendelian randomisation with IL6R genetic instruments as proxies for lifelong IL-6 receptor blockade (the pharmacology target of drugs such as tocilizumab). Primary analyses used UK Biobank (~486k adults, 11,643 with sepsis), with secondary cohorts including FinnGen and COVID-19 Host Genetics Initiative data.

Limitations

MR estimates lifelong genetic exposure, not a same-day ICU tocilizumab RCT for bacterial sepsis. Some secondary critical-care survival intervals are wide and compatible with no effect. Results support a therapeutic hypothesis and trial priority — they do not license off-label IL6RA use for every sepsis case.

How this study connects

Role on claims

Each row is a claim on a concept or method page where this paper supports, challenges, or qualifies the statement. Roles are hand-checked — not a model guess.

  • QualifiesSepsisconcept

    Blood leukocyte transcriptomes split ICU pneumonia sepsis into SRS1 vs SRS2 with different early mortality.

    In prospective UK ICU adults with community-acquired pneumonia sepsis, unsupervised transcriptomics defined SRS1 (~41%) as an immunosuppressed state with higher 14-day mortality than SRS2 (discovery HR 2.4; validation HR 2.8), classifiable with a seven-gene set.

    Scope note — different question — population genetic proxies for IL6R blockade, not ICU transcriptomic subtypes

    Limits the claim's scope: a different population, assay, or outcome.

  • SupportsSepsisconcept

    Genetically proxied IL6R blockade associates with lower sepsis risk and severity.

    Mendelian randomisation treating IL6R variants as lifelong proxies for receptor blockade found lower odds of sepsis in UK Biobank (OR 0.80) and stronger associations for critical-care sepsis phenotypes (e.g., OR 0.48), with a similar-sized protective association for severe COVID-19 where IL-6 blockade is already recommended.

    Evidence for the claim as stated.

  • SupportsSepsisconcept

    SRS classes describe acute ICU transcriptomic states after pneumonia sepsis; IL6R MR estimates lifelong pathway effects on sepsis risk across a volunteer biobank. Both can be true without being the same assay or the same treatment decision.

    Evidence for the claim as stated.

Open questions

Tensions this paper is part of

From concept pages' “where studies disagree.” Disagreement means the same question; scope means different assays, populations, or outcomes.

  • Scope difference — different assays, populations, or outcomes

    SupportsSepsis

    SRS classes describe acute ICU transcriptomic states after pneumonia sepsis; IL6R MR estimates lifelong pathway effects on sepsis risk across a volunteer biobank. Both can be true without being the same assay or the same treatment decision.

    Also on this tension

Related papers in this topic

Same topic cluster — not a recommendation engine.